Uni e si y o Sou he n Denma k
Dynamic Op ical Componen s by Combining Plasmonic Me asu aces wi h Piezoelec ic
MEMS
Vaagen Th ane, Paul Con ad
DOI:
10.21996/eac24207-cb4c-4b1e-a8ed-6a8 c3e717
Publica ion da e:
2025
Documen e sion:
Final published e sion
Documen license:
CC BY-NC-SA
Ci a ion o pulished e sion (APA):
Vaagen Th ane, P. C. (2025).
Dynamic Op ical Componen s by Combining Plasmonic Me asu aces wi h
Piezoelec ic MEMS
. [Ph.D. hesis, SDU]. Syddansk Uni e si e . De Tekniske Fakul e .
h ps://doi.o g/10.21996/eac24207-cb4c-4b1e-a8ed-6a8 c3e717
Go o publica ion en y in Uni e si y o Sou he n Denma k's Resea ch Po al
Te ms o use
This wo k is b ough o you by he Uni e si y o Sou he n Denma k.
Unless o he wise speci ied i has been sha ed acco ding o he e ms o sel -a chi ing.
I no o he license is s a ed, hese e ms apply:
• You may download his wo k o pe sonal use only.
• You may no u he dis ibu e he ma e ial o use i o any p o i -making ac i i y o comme cial gain
• You may eely dis ibu e he URL iden i ying his open access e sion
I you belie e ha his documen b eaches copy igh please con ac us p o iding de ails and we will in es iga e you claim.
Please di ec all enqui ies o [email p o ec ed]
Download da e: 22. Sep. 2025
UNIVERSITY OF SOUTHERN DENMARK
Doc o al Thesis
Dynamic Op ical Componen s
by Combining Plasmonic Me asu aces
wi h Piezoelec ic MEMS
Au ho :
Paul Con ad Vaagen Th ane
Supe iso :
P o . D . Se gey I. Bozhe olnyi
Co-supe iso :
D . Ch is ophe A. Di dal
A hesis submi ed in pa ial ul illmen o he equi emen s
o he deg ee o Doc o o Philosophy a he
Cen e o Nano Op ics
Mads Clausen Ins i u e
Feb ua y 2025
ii
Dynamic Op ical Componen s
by Combining Plasmonic Me asu aces wi h Piezoelec ic MEMS
Paul Con ad Vaagen Th ane
Feb ua y 2025
This wo k is licensed unde a C ea i e Commons
“A ibu ion-NonComme cial-Sha eAlike 4.0 In-
e na ional” license.
iii
Abs ac
Op ical me asu aces o e an inc edible weal h o oppo uni ies o con ol ligh
in new ways. Comp ised o sub-wa eleng h sized s uc u es hey enable e ec-
i e ma e ial p ope ies no p esen in na u ally occu ing subs ances, and
which can be ailo ed by design o con ol all p ope ies o ligh . Dynamic
me asu aces - me asu aces ha can adjus hei p ope ies du ing use - a e
especially p omising, and he e is cu en ly a la ge ongoing e o by he e-
sea ch communi y o iden i y and de elop hese de ices.
This hesis p esen s s udies on one such sys em. By eplacing he me al-
lic subs a e in gap su ace plasmon me asu aces wi h a piezoelec ic MEMS
mi o , we ealize me asu ace s uc u es coupled o a mic o-ca i y wi h a
a iable leng h. This esul s in a dynamic e lec i e me asu ace wi h as
esponse down o 5 µs, high e iciency o 50-90 %, ull 2πphase con ol, and
we demons a e he concep o se e al in e es ing use cases in nea IR e-
quencies. Th ough he in oduc o y chap e s and included a icles, he design,
ab ica ion, assembly and cha ac e iza ion o hese de ices is p esen ed, and
hei me i s and d awbacks a e discussed in ligh o o he de elopmen s in he
ield. Being based on wo ecen , bu by now well es ablished echnologies, he
inal conclusion is ha he pla o m ep esen s a sui able and cos -e ec i e
pla o m o ealizing dynamic me asu aces.
i
Resum´e
Op iske me a lade ilbyde en u olig igdom a mulighede o a kon olle e
lys p˚a nye m˚ade . Bes ˚aende a s uk u e i sub-bølgelængde s ø else muliggø
de e ek i e ma e ialegenskabe de ikke e il s ede i na u lig o ekommende
s o e , og som kan sk ædde sys ed design il a kon olle e alle lyse s egen-
skabe . Dynamiske me a lade - me a lade de kan jus e e de es egenskabe
unde b ug - e sæ lig lo ende, og de e i øjeblikke en s o igang æ ende
indsa s a o skningssam unde o a iden i ice e og ud ikle disse enhede .
Denne a handling p æsen e e s udie a e s˚adan sys em. Ved a e s a e
de me alliske subs a i gap su ace plasmon me a lade med e piezoelek isk
MEMS-spejl, ealise e i me a lade-s uk u e koble il en mik o-ka i e med
a iabel længde. De e esul e e i en dynamisk e lek e ende me a lade med
hu ig espons ned il 5 µs, høj e ek i i e p˚a 50-90 %, uld 2π asekon ol,
og i demons e e koncep e o le e in e essan e an endelses il ælde i næ
IR- ek ense . Gennem de indledende kapi le og inklude ede a ikle p æsen-
e es design, ems illing, samling og ka ak e ise ing a disse enhede , og de es
o dele og ulempe disku e es i lyse a and e ud iklinge p˚a om ˚ade . Base e
p˚a o nye e, men nu el e able ede eknologie , e den endelige konklusion, a
pla o men ep æsen e e en god og omkos ningse ek i pla o m il ealise -
ing a dynamiske me a lade .
P e ace and Acknowledgemen s
This hesis is he esul o close collabo a ion be ween he esea ch g oups a
he Cen e o Nano Op ics a he Uni e si y o Sou he n Denma k and he
Mic o-Op ics g oup a SINTEF in No way, espec i ely led by my supe iso
o his wo k, P o . D . Se gey I. Bozhe olnyi, and my co-supe iso D .
Ch is ophe A. Di dal. Du ing he cou se o his wo k I ha e also isi ed P o .
D . Ola Solgaa d, a he Ginz on Labo a o y, S an o d Uni e si y.
This wo k has ecei ed unding om he Resea ch Council o No way un-
de p ojec numbe 323322; he ATTRACT p og am (Eu opean Union Ho i-
zon 2020 Resea ch and Inno a ion P og am 101004462); and he Eu opean
Inno a ion Council unde p ojec numbe 101185769.
The wo k p esen ed in his hesis would in no way ha e been possible
alone, and he e a e a lo o people I am g a e ul o and who ha e made
i all possible. Fi s o all, my supe iso Se gey is an expe a e ocusing
o e ambi ious s uden s on he mos impo an asks a hand. I am a aid I
ha e no always made his job easy, bu I can since ely say I am immensely
g a e ul o all his insigh ul sugges ions, clea guidance, suppo and kindness.
Being o he mos pa loca ed on opposi e sides o Skage ak has complica ed
some hings, and I hugely app ecia e he la ge e o Se gey has made o make
me an in eg a ed pa o his g oup none heless. They a e a e y welcoming
g oup and I ha e had he g ea pleasu e o collabo a ing wi h many o i s
membe s, I hank hem all and hope we will keep he collabo a ion going.
I am immensely g a e ul o my local supe iso oo, Ch is ophe , who has
likewise made a lo o accommoda ions o acili a e he p ojec . His en hu-
siasm, mo i a ional suppo and o ganiza ional skills ha e somehow made i
possible o juggle he many asks wi hin his p ojec and o he s. In ex ension
o him, I am also hank ul o SINTEF o le ing me ha e his oppo uni y.
My g a i ude u he mo e goes ou o Ola and his s uden s a S an o d,
a g oup o highly in elligen people bes cha ac e ized by hei kindness and
inclusi eness. Ola addi ionally has a knack o spa king in e es and ideas,
and I g ea ly alue he ime he de o es o help us s uden s wi h ou p ojec s.
The idea o his p ojec would ne e ha e exis ed i i had no been o
Tho Bakke and he o he people in ol ed in de eloping he MEMS we ha e
been using, which includes many o my an as ic colleagues a SINTEF. Some
o my colleagues ha e no been in ol ed, bu hey a e an as ic oo and I
i
hank hem o making my wo kplace g ea . My colleague and iend Runa is
especially awesome, and I wan o hank him o ha , as well as his wi e Line
o being awesomes .
My amily and many g ea iends also dese e a men ion, wi hou hem
his wo k migh ha e been possible - bu no by me. I hank hem wi h all my
hea . I I ge he impossible ask o naming jus a ew, I am o e e g a e ul
o Anne-Line, H˚akon, Ma ie and K is ian, o whom I know I can always go
o suppo - o jus a s upidly long ski excu sion i ha is wha I need.
Las bu ce ainly no leas , I would like o hank Chao Meng, who I ha e
had he g ea pleasu e o collabo a ing wi h hese las yea s and who I conside
a close iend. I g ea ly alue his immense kindness, and admi e his b illiance
o expe imen al wo k which has made his p ojec possible.
Con en s
Abs ac iii
Resum´e i
P e ace and Acknowledgemen s
1 In oduc ion 1
1.1 Mo i a ion ............................. 1
1.2 Aims o his hesis ......................... 2
1.3 Thesis ou line ............................ 3
2 S a ic me asu aces 5
2.1 O e iew and s a e o he a ................... 5
2.1.1 The ansi ion o la op ics ................ 5
2.1.2 Ca ego ies o me asu aces ................. 7
2.1.3 Applica ions ........................ 16
2.1.4 Challenges ......................... 17
2.2 Me asu ace design ......................... 18
2.3 Me asu ace ab ica ion ...................... 23
3 Ac i e me asu aces 27
3.1 O e iew and s a e o he a ................... 27
3.1.1 MEMS me asu aces .................... 29
3.2 Piezoelec ic MEMS ........................ 31
3.3 Combining piezoelec ic MEMS and me asu aces ........ 35
3.4 MEMS me asu aces, a compa ison ................ 46
4 Conclusion and ou look 51
ii
iii CONTENTS
5 A icles 53
5.1 Dynamic piezoelec ic MEMS-based op ical me asu aces . . . 53
5.2 Full- ange bi e ingence con ol wi h piezoelec ic MEMS-based
me asu aces ............................ 66
5.3 MEMS Tunable Me asu aces Based on Gap Plasmon o Fab y–P´e o
Resonances ............................. 74
5.4 MEMS- unable opological bilaye me asu aces o econ ig-
u able dual-s a e phase con ol .................. 82
5.5 Me asu ace Pola ime e o S uc u al Imaging and Tissue Di-
agnos ics ............................... 94
Re e ences 113
2.1. OVERVIEW AND STATE OF THE ART 7
Figu e 2.1: aA p ism can be used o shi he di ec ion o an op ical beam
o ligh . bA blazed g a ing - a ype o di ac i e op ical elemen - can do he
same, bu w apping he phase by 2πenables conside able minia u iza ion. c
Op ical me asu aces can be used o make di ac i e op ical elemen s, bu a e
no limi ed o his.
coun less design choices like combina ions o ma e ials and geome ies - only
limi ed by you imagina ion and you ab ica ion echniques - ansla e in o
a huge a ie y o di e en beha io s in he e ec i e ma e ial. The e ha e
been demons a ed nume ous ways o con olling bo h phase and ampli ude,
and by in oducing aniso opic geome ies i is ela i ely s aigh o wa d o do
his independen ly o o hogonal pola iza ion s a es. As a speci ic example,
me asu ace based di ac i e lenses ha e al eady shown be e nume ical ape -
u e, e iciency and pola iza ion con ol han mo e adi ional app oaches [11],
and mo e p og ess is expec ed on con ol o ch oma ic e ec s and unabili y.
Bu how do hey wo k? The ollowing sec ion will gi e a b ie o e iew o
di e en ypes o me asu aces and hei ope a ional p inciples.
2.1.2 Ca ego ies o me asu aces
The deg ees o eedom in ma e ials, geome ies, a ge ed wa eleng hs and
all he a ious con igu a ions o di e en applica ions makes classi ica ion o
me asu aces a challenge, bu he e a e some main ca ego ies ha gi e a good
o e iew o mos o he concep s used o da e [12–16]. Pe haps he mos
ins uc i e way o ca ego izing is by how he me asu ace impa s a phase
delay o he elec omagne ic wa e. Fo a adi ional lens his is done by
a ying he pa h leng h di e en ays o ligh ha e o a e se h ough he
lens, aking in o accoun how he ligh e ac s a he su aces. As we ha e
men ioned egula di ac i e op ics does he same bu w aps he phase by 2π.
Wi h me asu aces i is likewise possible o change he phase by a ying he
8CHAPTER 2. STATIC METASURFACES
op ical pa h dis ance, bu ins ead o a ying he physical dis ance h ough he
ma e ial his is done by a ying he local e ec i e e ac i e index seen by he
p opaga ing ligh . Ano he echnique is o make use o he phase delay ha
comes in connec ion wi h a esonance, we shall b ie ly discuss bo h o hese
op ions a e looking a a hi d way o con ol he phase.
Geome ic phase
When con e ing one s a e o pola iza ion in o ano he he e is a con ibu ion
o he phase change which is dependen on how he pola iza ion is changed
called he geome ic phase, o o en he Pancha a nam-Be y phase. I a ises
due o he geome y, and mo e speci ically he cu a u e, o he pa ame e
space desc ibing he deg ees o eedom - which o pola iza ion can be ep e-
sen ed by he Poinca ´e sphe e [17]. As such, i is no limi ed o pola iza ion
e ec s bu also appea s in o he a eas like he spin o an elec on which sim-
ila ly can be desc ibed by a Bloch sphe e, while also being ele an o mo e
complica ed sys ems [18]. To illus a e he mos used example, we look a
ansmission h ough a bi e ingen ma e ial, wi h o dina y and ex ao dina y
axes aligned in he o hogonal ˆxand ˆydi ec ions as desc ibed by a Jones
ma ix
J(0) = 1 0
0eiδ,
whe e we assume pe ec ansmission ampli ude, and igno e he dynamic
phase change by pulling ou eiδxand se δ=δy−δxbeing he phase di e ence
be ween he wo linea eigenpola iza ion s a es upon ansmission. Following
he same p ocedu e as o example in [19] we can o a e he basis by an angle
θand change o a ci cula pola iza ion basis wi h uni ec o s ˆ
l= (ˆx+iˆy)/√2
and ˆ = (ˆx−iˆy)/√2 gi ing a new Jones ma ix
˜
J(θ) = 1
2(1 + eiδ)1 0
0 1+1
2(1 −eiδ)0e−i2θ
ei2θ0.(2.1)
F om his equa ion i becomes immedia ely appa en ha i he inciden ligh
is ci cula ly pola ized, and we design he ma e ial o ac as a hal -wa epla e
wi h δ=π, hen we can impa a phase o 2θ o he c oss-pola ized ligh by
a ying he amoun o o a ion θ. The use ulness o his comes om he ac
ha wi h me asu aces i is possible o a y he di ec ion o he bi e ingence
locally ac oss a su ace as indica ed in Figu e 2.2. This way, a blazed g a ing
can o example be ealized by g adually a ying θalong one di ec ion o he
su ace.
2.1. OVERVIEW AND STATE OF THE ART 9
Figu e 2.2: aand b, a pola iza ion s a e going h ough wo elemen s wi h
eigenpola iza ion di ec ion o a ed compa ed o each o he . In bo h cases he
pola iza ion is changed om ci cula o he opposi e handedness, bu he e is
a di e ence in he phase o he ansmi ed ligh which is dependen on he
o ien a ion o he eigenpola iza ion axes. cSEM image o a geome ic phase
me asu ace made by Ch is ophe Di dal a SINTEF.
No e ha he wo o hogonal ci cula pola iza ion s a es acqui e opposi e
phases, and hese ypes o me asu aces a e he e o e inhe en ly pola iza ion
dependen . Fo some applica ions his can be used as an ad an age, while o
o he applica ions i migh mean a educed e iciency. Fo example, in [20]
he au ho s demons a e a lens which is ocusing o one helici y while being
de ocusing o he o hogonal one. Ad an ages wi h his echnique include
he ac ha he phase is no as di ec ly wa eleng h dependen as wi h some
o he o he e ec s, and migh he e o e be mo e b oadband. Fu he mo e,
hese me asu aces o en only need one nanos uc u e geome y wi h an e en
ill ac o , which can make i easie o op imize ab ica ion and ge high phase
esolu ion wi h e y e en ampli ude.
In he abo e discussion we assumed pe ec ansmission and chose δ=π,
howe e , i is possible o a y he bi e ingence o con ol bo h phase and
ampli ude. I assuming only one ci cula pola iza ion s a e is inciden and
il e ing ou any ligh no con e ed o he o hogonal pola iza ion a e wa ds,
Equa ion (2.1) ells us ha he ampli ude o he ansmi ed ligh will be
dependen on he deg ee o bi e ingence and he phase will be a combina ion
o he dynamic and geome ic phase, his has been demons a ed in [21]. I
is wo h no ing ha one can also use ellip ic o ci cula eigenpola iza ion
s a es, which makes i possible o independen ly con ol ellip ic o linea s a es
10 CHAPTER 2. STATIC METASURFACES
Figu e 2.3: aIn-plane iew o silicon pilla s ha unc ion as an e ec i e
medium o ligh wi h wa eleng h 1.55 µm. bTop iew o simila silicon
pilla s - he a ying diame e o he pilla s changes he e ec i e e ac i e
index and hus he phase delay upon ansmission. Me asu aces made by
Ch is ophe Di dal a SINTEF.
o pola iza ion [17]. Howe e , his can equi e mo e complica ed uni cell
s uc u es [22,23]. Ano he echnique ha simila ly gi es con ol o e gene al
pola iza ion s a es bu wi h po en ially simple ab ica ion, is by combining
he geome ic phase wi h o he phase con ol echniques like he unca ed
wa eguide me hod [24].
T unca ed wa eguides
Figu e 2.3 is an illus a ion o dielec ic pilla s p o uding om a su ace. The
phase shi o ligh p opaga ing h ough such a me asu ace can be unde s ood
as p opaga ion h ough a media wi h e ec i e e ac i e index in be ween hose
o he pilla s and su ounding media
ϕ=2π
λne h,
wi h hbeing he heigh o he pilla s and ne he e ec i e e ac i e index.
Keeping he heigh o all pilla s he same o simpli y ab ica ion, a ia ion
o he phase can hen be achie ed by al e ing he diame e o shape o he
pilla s o change ne . This e ec i e e ac i e index can be es ima ed as he
undamen al mode index o a single s ep-index ci cula wa eguide model [25],
2.1. OVERVIEW AND STATE OF THE ART 11
wi h he la ges e o s being due o mul iple e lec ions a bo h ends o he
pilla [26].
Fo no mal incidence and wi h iso opic s uc u es such me asu aces will
be pola iza ion independen , bu i is s aigh o wa d o make pola iza ion de-
penden componen s by o example using ellip ic o ec angula c oss-sec ions,
which can be combined wi h he geome ic phase o con ol gene al pola iza-
ion s a es [24]. Whe he hey a e pola iza ion dependen o independen , he
phase is di ec ly dependen on he wa eleng h as well as being sensi i e o he
incidence angle, and mos o hese me asu aces ha e been designed o sin-
gle wa eleng h applica ions. None heless, he e a e many di e en echniques
being pu sued o inc ease he bandwid h o ge con ol o e he spec al and
angula dispe sion o o he pu poses [27]. A couple examples include inc eas-
ing he complexi y o he uni cell wi h mul iple o mo e complica ed pilla s,
gi ing mo e pa ame e s o adjus independen ly [28,29], adding a e lec i e
back su ace [30] and cascading se e al me asu aces [31]. Fo he unca ed
wa eguide me hod in gene al, ad an ages include high e iciency when using
lossless dielec ic ma e ials, and localized modes ha educe c oss- alk be-
ween neighbo ing s uc u es and simpli ies design. On he o he hand, a
d awback is he ela i ely all s uc u es ha mus be made, oughly a ound
one wa eleng h depending on he e ac i e index con as be ween he pilla
and su ounding ma e ial. This leads o aspec a ios and a ying ill ac o s
ha can be challenging o ab ica e, especially when wan ing o amp up p o-
duc ion pas single p o o ypes. I is also possible o use dielec ic s uc u es
wi h smalle aspec a ios, hese do no unc ion he same way bu a he make
use o esonances [26].
Resonan phase
As physicis s o en do [32] we shall con empla e he o ced ha monic oscilla o
¨x+ Γ ˙x+ω2
0x=1
mF( ),
wi h Γ being a damping e m. Assuming a ime-ha monic beha io , wi h he
same equency as he d i ing o ce we can w i e
(−ω2+iΓω+ω2
0)x0eiω =1
mF0eiω ,
which can be sligh ly ew i en o emphasize ha he ine ial, damping and
es o ing e ms ha e di e en phases wi h espec o he d i ing o ce
x=x0eiω =F0eiω /m
eiπω2+eiπ/2Γω+ω2
0
.(2.2)
12 CHAPTER 2. STATIC METASURFACES
Clea ly, when d i ing he oscilla o way below he esonance equency ω≪ω0
he ω2
0 e m domina es (we a e assuming he damping e m is small), which
means he oscilla ions a e in phase wi h he d i ing o ce. Meanwhile, a he
esonance he ine ial and es o ing e ms cancel, delaying he oscilla ion phase
by a qua e o a pe iod, which u ns o hal a pe iod o ω≫ω0when he
d i ing o ce is so as ha he ine ial e m domina es. These conside a ions
s ill apply when aking in o accoun he adia ion eac ion o ce by adding a
damping e m p opo ional o ...
x[12,33,34]. I we ake Equa ion 2.2 o be
desc ibing a cha ge in a po en ial being d i en by an applied elec ical ield, we
can use i o ind he pola iza ion ield which gi es us he well known Lo en z
model o he dielec ic esponse - see o example [35]. In his case, howe e , we
a e in e es ed in his because i sugges s ano he way o design a me asu ace:
Make a esonan sys em whe e you can a y ω0, o a ce ain ω he phase o
ligh sca e ed by his esonance can hen be con olled. Fo example, a hin
me allic ba being hi by an EM wa e pola ized along i ’s long axis will ha e
an elec ic dipole esonance wi h esonan wa eleng h abou wice he leng h
o he ba [13], and adjus ing his leng h will he e o e change he phase o he
sca e ed ligh . No e ha his es ima e is no accu a e a op ical equencies,
whe e me als can no longe be conside ed as pe ec conduc o s. The p oblem
wi h his simple app oach is ha i only allows con olling he phase o wi hin
a alue o π, while ull con ol o he sca e ed ield equi es a phase ange o
2π.
A couple main echniques ha e been de eloped o achie e ull phase ange
and high e iciency. Fi s ly, dielec ic pa icles wi h a high e ac i e index and
a ans e se size app oxima ely equal o he wa eleng h in ha ma e ial can be
designed o ha e o e lapping elec ic and magne ic esonances which ha e he
same ampli ude and a e in phase. This condi ion, known as he i s Ke ke
condi ion [36], esul s in cancella ion o he back-sca e ed ield enabling high
ansmission e iciency in addi ion o he 2πphase shi . Ye , when a ying
he size/shape o hese s uc u es o change he phase he ansmission is
a ec ed [37,38], because he dipole esonances a e no longe pe ec ly o e -
lapped, in addi ion o coupling be ween neighbo ing s uc u es a ec ing he
o al esponse [39]. Fo hese easons he unca ed wa eguide app oach e-
sul s in highe e iciency while his esonan app oach gi es mo e enhanced
ields, and can be easie o ab ica e as he aspec a ios a e no as la ge [40].
Meanwhile, o phase con olling plasmonic me asu aces we can ca ego ize
hem oughly in wo me hods, one ha can be used o ansmission and one
ha wo ks in e lec ion. The ansmissi e ones combine esonan e ec s wi h
he geome ic phase o ge ull phase co e age. Two common con igu a ions
a e using -shaped an ennas o spli ing esona o s, wi h one symme ic mode
2.1. OVERVIEW AND STATE OF THE ART 13
Figu e 2.4: aIllus a ion o a me asu ace comp ised o a e lec i e subs a e,
a dielec ic space and plasmonic nanoan ennas o a ying sizes. bSimula ion
o a gap su ace plasmon esonance, yellow a eas a e gold, while he es is
ai . The blue indica es a eas wi h no ield enhancemen while ed indica es
la ge ield enhancemen . cSEM image o a pola iza ion independen blazed
g a ing me asu ace. dSEM image o a pola iza ion dependen blazed g a ing
me asu ace, no ice he aniso opic s uc u es. The scale ba is o bo h SEM
images. The me asu aces we e made by Chao Meng a SDU.
and one an isymme ic mode, and whe e he asymme ic geome y allows he
an isymme ic mode o couple wi h inciden ligh [41]. Al hough hey a e sim-
ple o ab ica e, being dependen on he geome ic phase hey a e pola iza ion
dependen and he e iciency is dic a ed by he c oss-pola iza ion e iciency,
which is limi ed o hese kinds o s uc u es - expe imen al demons a ions
a e ha e had a ound 10 % e iciency while heo e ically he limi is 25 % la gely
due o symme ic o wa d and backwa ds sca e ing [42–46]. Mo e complica ed
con igu a ions such as dielec ic s uc u es co e ed wi h a hin laye o me al
deposi ed on op ha e shown highe e iciency up o a ound 40 %, wi h la ge
e iciencies being possible due o canceling he back-sca e ing using se e al es-
onances [42,43]. A simple and e icien way o b eaking he o wa d/backwa d
symme y is o add a e lec i e su ace, in which case you me asu ace ob i-
ously will wo k in e lec ion, b inging us o gap su ace plasmon me asu aces.
Gap su ace plasmon me asu aces
These me asu aces consis o a small plasmonic an enna sepa a ed om a
g ound plane by a dielec ic space , see Figu e 2.4. When he dielec ic space
14 CHAPTER 2. STATIC METASURFACES
is hin enough, su ace plasmon pola i ons [47] on he wo me al-dielec ic in-
e aces can in e ac h ough nea ield coupling [48]. As a esul , he sys em
ac s as a Fab y-P´e o ca i y o he gap su ace plasmon modes wi h he ca i y
size being de e mined by he leng h o he an enna [49]. In p inciple highe
o de modes a e possible, bu we shall s ick o he lowes o de an isymme -
ic mode, whe e he cu en s in he an enna and g ound plane go in opposi e
di ec ions and he sys em ac s as a magne ic dipole. I he pa ame e s a e
chosen app op ia ely, see Sec ion 2.2, his esul s in a deeply sub-wa eleng h
esona o ha can be easily uned by adjus ing he leng h o he an enna,
which has a s ong sca e ing c oss-sec ion and a la ge phase ange [7,48,
50]. This sys em can be used wi h he geome ic phase, bu in con as o he
ansmissi e plasmonic s uc u es i does no need geome ic phase o achie e
ull 2πphase ange. One way o unde s anding his is ha away om he
esonance he sys em ac s as a mi o , applying a πphase shi o he e-
lec ed elec ic ield, while a esonance he sys em ac s as a magne ic dipole
and hus as a magne ic mi o - applying he πphase shi o he magne ic
ield and hus he elec ic ield is e lec ed wi h ze o phase. Con inuing pas
he esonance will con inue changing he phase un il he elec ic ield is again
e lec ed wi h πphase [13,51]. In p ac ice howe e , some o he phase ange
is usually sac i iced in o de o ha e less abso p ion losses a he esonance,
his ade-o is done by adjus ing he dielec ic laye hickness and has he
added ad an age o b oadening he esonance, he eby simpli ying ab ica ion
as he size ole ances become la ge . Signi ican ly mo e han 2πphase ange is
also possible, by in oducing se e al de uned esona o s o each uni cell [52].
Fu he mo e, i is s aigh o wa d o con ol he phase independen ly o o -
hogonal linea pola iza ion s a es, as we shall demons a e in Sec ion 2.2,
by adjus ing he wo in-plane dimensions o he an enna, ellip ic o ci cu-
la eigenpola iza ion s a es a e also possible wi h mo e complica ed uni cell
s uc u es [53]. We should no e ha o la ge sepa a ions be ween he g ound
plane and plasmonic an enna he nea ield coupling ceases and ins ead we ge
a Fab y-P´e o esonance be ween he g ound plane on one side, and he me a-
su ace nanos uc u es on he o he [48,54]. This will be discussed in mo e
de ail in Chap e 3, a comp ehensi e e iew is also ound in [55]. I is also
wo h no ing he analogy o e lec a ays [56–58], which sha e many cha ac-
e is ics bu ha e some implemen a ions which a e no cu en ly easible on
he nanoscale (and a op ical equencies) such as connec ing s uc u es wi h
indi idually unable a ac o s [59].
Al hough hey a e limi ed o wo king in e lec ion, he e a e se e al ad an-
ages wi h gap su ace plasmon me asu aces. O he han simple ab ica ion,
hey can be b oadband and e icien , an example being 80% e iciency and
2.1. OVERVIEW AND STATE OF THE ART 15
150 nm bandwi h a 850 nm using gold [50], in addi ion o wo king o a la ge
ange o incidence angles. The bandwid h can also be ex ended by in oduc-
ing mo e complica ed esona o s, a he expense o e iciency [60]. The losses
a e mainly due o abso p ion, and can be used o con ol ampli ude in addi-
ion o phase, o o make na owband as well as b oadband abso be s [48].
By using gold, he me asu aces can also bene i om es ablished me hods o
bio unc ionaliza ion o make senso s [61].
Ma e ial-wise, hough gold is sui able o ab ica ion, biosensing and i s
chemical s abili y, i does howe e ha e a lo o abso p ion in he isible e-
quencies and consequen ly he e iciency su e s a hese wa eleng hs. Sil e
is a good low-loss op ion [62], bu lacks chemical s abili y and like gold is no
compa ible wi h CMOS p ocessing. Aluminium is a CMOS compa ible al e -
na i e [63] wi h mo e losses han sil e , bu less losses han gold a sho e
equencies, and is chemically p o ec ed by he na u ally o ming oxide laye .
Finally, o he possibili ies a e also being de eloped, such as conduc i e ox-
ides [64]. Fo example, i anium ni ide is CMOS compa ible and has been
used o make plasmonic nanos uc u es wi h cha ac e is ics simila o hose
made o gold bu wi h lowe losses when using wa eleng hs a ound 500 nm [65].
O he kinds o me asu aces
The abo e men ioned ca ego ies ocus on how phase is impa ed, while i is
also possible o make me asu aces ha do no aim o con ol he phase bu
a he ocus on he ampli ude o he e lec ed o ansmi ed ligh . A well
known analogy o mo e adi ional di ac i e op ics is he F esnel bina y zone
pla e whe e, ins ead o a a ying phase p o ile, a bina y g a ing is used o make
a ocusing lens by selec i ely le ing h ough pa s o a beam ha will in e e e
cons uc i ely a he desi ed ocal spo . In [66] his p inciple is implemen ed
o make an ach oma ic me asu ace lens o isible ligh . This is accomplished
by cascading h ee sepa a e me asu ace laye s, each laye consis ing o an
a ay o plasmonic nanodisks. In addi ion o sepa a ely a ying he nanodisk
diame e s and spacings, he au ho s also use di e en ma e ials (gold, sil e
and aluminium) in each laye o op imize he combined lens o ed, g een and
blue wa eleng hs. Ano he example is ound in [67] whe e an a ay o gold
disks a e sepa a ed om a g ound plane by a hin dielec ic space , bu ins ead
o being op imized o phase con ol he s uc u es a e ailo ed o abso p ion.
By delibe a ely analyzing and aking in o accoun he amoun o damping
in hei ma e ials, he au ho s achie e nea pe ec abso bance o e a b oad
ange o angles and wi h minimal dependence on pola iza ion. No ably, he
abso bance is e y sensi i e o he e ac i e index o he su ounding dielec ic
16 CHAPTER 2. STATIC METASURFACES
media, and he sys em is he e o e demons a ed as a plasmonic senso .
Ano he ype o me asu ace ha does no i he ca ego ies abo e is wha
can be classi ied as a nonlocal me asu ace. So a we ha e only discussed
a ays o s uc u es whe e he esponse is local, which is o say ha he e-
sponse is only de e mined by he s uc u es in he immedia e icini y - o en
wi h minimal c oss- alk be ween neighbo ing s uc u es. In many cases his
is on pu pose as i makes design simple . Howe e , sys ems whe e he e sca -
e ing a one loca ion is dependen on ields and s uc u es o e a b oad a ea
can gi e a high deg ee o con ol o e he spec al and angula p ope ies [68].
Fu he mo e, such me asu aces can be u ilized o g ea ly inc ease he e ec i e
in e ac ion leng h be ween ma e and ligh passing h ough a hin me asu -
ace, which can be used o inc ease nonlinea e ec s [69]. Thus, me asu aces
can be used no only o change he phase, ampli ude and pola iza ion o ligh ,
bu also he wa eleng h. I should be no ed ha nonlinea e ec s can also be
enhanced in o he ways such as h ough ield enhancemen [70,71].
2.1.3 Applica ions
The numbe o possible applica ions o op ical me asu aces is seemingly end-
less, wi h an e e g owing lis o p omising al e na i es [4,14,15]. No exhaus-
i e o e iew shall be a emp ed he e, bu a he a ew selec ed opics wi h
some examples will be men ioned o gi e an imp ession o he b ead h o he
esea ch ield.
Me asu ace lenses [28,72] and ela ed componen s like axicons [28] en-
able minia u iza ion o op ical sys ems which is o g ea impo ance o bo h
imaging and p ojec ion sys ems o consume elec onics, bu also highe -end
ma ke s such as endoscopes [73]. Fo eme ging applica ions like augmen ed
eali y, o -axis op ical componen s a e especially in e es ing as i allows o
combine hese p ojec ion and imaging sys ems seamlessly and unob usi ely
in o he line o sigh [74,75]. Pa e n p ojec ion is a simila high- olume
ma ke whe e me asu aces a e al eady being used in do -p ojec ion o 3D-
measu emen s comme cially [3] and wi h u he op imiza ion al eady demon-
s a ed [76]. Mo e gene ally o cou se, me asu aces a e no limi ed o p o-
jec ing do s and one can imagine much mo e sophis ica ed echniques u ilizing
all he esea ch done on holog am gene a ion [77,78]. Toge he wi h beam
s ee ing, o example o lida [79], hese a eas a e expec ed o accoun o he
majo i y o he ma ke in he nea u u e [3].
Senso s a e ano he big ca ego y, and we ha e al eady discussed he plas-
monic e ac i e index senso in [67]. Ano he example can be ound in [80],
whe e a hype spec al sys em o biosensing is p esen ed. The scheme con-
2.3. METASURFACE FABRICATION 23
Figu e 2.9: aDa k ield mic oscope image o he exposed and de eloped esis
o a se ies o dose es s, wi h close up da k ield and clea ield images in b
and c. No e ha his is an ea lie design han ha shown in Figu es 2.8,2.10
and 2.11, and has an ellip ic ape u e.
2.3 Me asu ace ab ica ion
The ab ica ion p ocess s a s wi h a die om a polished silicon wa e , on
which a 3 nm i anium, 120 nm gold, 3 nm i anium, and 60 nm SiO2a e de-
posi ed using e apo a ion (To nado 400 C yo ox). The i anium laye s a e o
ensu e good adhesion. Subsequen ly, a 100 nm laye o poly(me hyl me hac y-
la e) (PMMA A2, Mic oChem) is deposi ed using spin coa ing. This PMMA
laye unc ions as elec on beam esis and is pa e ned wi h a SEM (JEOL
JSM-6500F ield-emission SEM wi h a Rai h Elphy Quan um li hog aphy sys-
em), be o e being de eloped. A e his p ocess, he PMMA should co e
he subs a e e e ywhe e excep o a he loca ions whe e he nanob icks a e
in ended o be.
No e ha when combining he me asu aces wi h MEMS as desc ibed in
Chap e 3, he nanos uc u es a e ab ica ed on a glass subs a e (Bo o loa
33, Wa e Uni e se), and he PMMA is coa ed di ec ly on he glass. Since
he e is no conduc i e gold laye in his case, o p e en cha ge buildup du ing
exposu e a 40 nm hick conduc i e polyme laye (AR-PC 5090, All esis ) is
spin coa ed on op o he PMMA p io o e-beam exposu e and emo ed p io
o PMMA de elopemen by insing in DI wa e .
Following de elopmen , he s uc u es a e inspec ed in an op ical mic o-
scope, see Figu e 2.9, be o e ano he 3 nm i anium adhesion laye and 50 nm
gold a e deposi ed (To nado 400 C yo ox). Finally, he chip is subme ged
o e nigh in ace one which desol es he PMMA and comple es he li -o p o-
cess. SEM images o esul ing s uc u es a e shown in Figu e 2.10.
24 CHAPTER 2. STATIC METASURFACES
Figu e 2.10: SEM images o me asu ace lenses o collima ion o a ibe ou -
pu . aMe asu ace lens wi h diame e 30 µ.bClose up image o s uc u es
in a. cMe asu ace lens wi h unsuccess ul li -o , la ge pa s o he ci cle is
co e ed by connec ed gold (b igh a eas).
Figu e 2.11 is an image o he collima ing me asu ace lens being es ed.
The samples we e es ed using a supe -con inuum lase wi h a a iable il e
(Supe K Ex eme, NKT Pho onics). The angula dispe sion was measu ed
o be a ound 0.05 deg ees pe nm as expec ed om Equa ion 2.3, while he
di e gence o he esul ing beam was es ima ed o 1 deg ee by compa ing
wi h he ou pu om a ibe wi h known di e gence as seen in he Fou ie
image plane. Addi ionally, one sample was es ed wi h a 100 s, 250 mW
lase o 20 minu es wi hou showing any signs o deg ading. Al hough his
p elimina y s udy shows g ea p omise o enabling a minia u e sys em o
empo al ocusing in 2-pho on mic oscopes, o a nex i e a ion i would be
ad an ageous o inc ease he dispe sion by using a me asu ace whe e one can
con ol he ch oma ici y such as demons a ed in [30], con e sely, he same
ype o me asu ace could be used o educe angula dispe sion o applica ions
whe e i is undesi able.
2.3. METASURFACE FABRICATION 25
Figu e 2.11: One o se e al me asu ace lenses is illumina ed a an angle by
a single mode ibe (da k shape p o uding om he le ). The illumina ed
me asu ace collima es he ligh and e lec s i in he no mal di ec ion owa ds
he came a (b igh ci cle), he unillumina ed me asu aces can be seen as da k
ci cles. A low in ensi y and wide beam is used o ligh up he ield o iew.
An alignmen ma ke (c oss) can be seen a he bo om o he pic u e.
26 CHAPTER 2. STATIC METASURFACES
Chap e 3
Ac i e me asu aces
3.1 O e iew and s a e o he a
Fo passi e me asu aces he op ical beha io is locked a he ime o ab ica-
ion. Ac i e me asu aces, on he o he hand, ha e he abili y o adjus he
op ical esponse a some la e poin . This change can be con olled on pu pose
by an ex e nal sys em, o i could occu as a esul o some physical p ocess
o which he me asu ace is designed o measu e o eac o. The las case is
e iden ly use ul o sensing pu poses, while he o me has many echnological
applica ions such as a i ocal lenses, mul iplexing, beam s ee ing and pola -
iza ion con ol. The la ge in e es and pa icipa ion in he apidly expanding
ield is he e o e o no su p ise, and he e ha e been emendous amoun s o
de elopmen s o e he las couple yea s [53,103–105].
To achie e dynamic unc ionali y, we s a om he basic p inciples o s a ic
me asu aces bu inco po a e some cha ac e is ic ha can be al e ed pos -
p oduc ion. An example could be how a localized su ace plasmon esonance
changes equency as he e ac i e index su ounding he an enna is al e ed,
which could be due o he p esence o some measu and o by con olling he
p ope ies o he media in some o he way. Ano he al e na i e is o al e he
geome y o he sys em, no ably, his is he me hod used o he a icles in
Chap e 5. As he eade by now su ely is ge ing accus omed o, we shall
summa ize his ield in a b ie manne wi h a limi ed se o examples, by
ouching upon some o he main p inciples used o acqui e unabili y, while
sa ing MEMS-based sys ems o he nex sec ion.
I you wan o ac i ely swi ch a e ac i e index, liquid c ys als a e a g ea
27
28 CHAPTER 3. ACTIVE METASURFACES
op ion. They consis o aniso opic molecules ha can be aligned and con-
olled by applying an ex e nal ol age, which changes he e ac i e index o
he ma e ial (by changing he o ien a ion o he o dina y and ex ao dina y
op ical axes). This p inciple can be used o make unable me asu aces unc-
ioning bo h in ansmission and e lec ion, as well as o con olling phase and
ampli ude [106,107], and has e en been used o unable nonlinea e ec s [108].
Ad an ages include he la ge op ical e iciency and change in e ac i e index
ha a e possible, as well as a ma u e and eliable echnology pla o m. The
main disad an age wi h liquid c ys als is ha he ime i akes o eo ien he
molecules limi s swi ching speed o a ound 1 kHz [105], al hough esea ch in o
new ypes o molecules could imp o e his o some deg ee [109].
Ano he e ec wi h la ge e ac i e index change is he swi ching be ween
c ys al and amo phous phases in some ma e ials such as GeSbTe (GST) and
VO2Fo GST, o example, his is accompanied by a 50 % educ ion o he
e ac i e index, depending on he wa eleng h. This is al eady used in op-
ical memo y s o age a eally as ope a ion speeds. Howe e , i is u ning
ou o be di icul o ealize he same ew i ing speeds o me asu aces, as
i is no possible o include he same p o ec i e laye s and elec odes [110].
None heless, high quali y me asu aces wi h ∼70 −80 % e iciency ha e been
demons a ed o bo h ma e ials, and o swi ching speeds on he o de o
1 kHz [111,112]. I is wo h no ing ha he swi ching om c ys al o amo -
phous s a e is usually much as e han opposi e, and can be used o example
o op ical limi ing in high-powe applica ions [113].
In some noncen osymme ic c ys als he e ac i e index can be al e ed in
p opo ion o an applied elec ic ield. Called he Pockels e ec , his change
only equi es a sligh displacemen o he ions in he c ys al la ice and can be
used o modula e he e ac i e index wi h swi ching speeds well in o he GHz
ange [114]. The ca ch, howe e , is ha hese ma e ials, o example LiNbO3,
PZT and BaTiO3a e challenging o use in nano ab ica ion p ocesses, and
he ypical e ac i e index change is small. High speed demons a ions a e
he e o e o en based on sho spec al shi s o na owband esonances o
ampli ude modula ion [115]. A u he al e na i e o such high-speed de ices
is using elec ical doping, also e e ed o as plasma dispe sion e ec o D ude
e ec [114,115], which can be used o al e he plasma equency o conduc i e
oxides by changing he densi y o ee cha ge ca ie s [116,117]. The e ec can
also be used o adjus he p ope ies o 2D ma e ials such as g aphene [61].
In bo h cases he e ec can be s ong, bu in bo h cases he e ec i eness o
hese sys ems is limi ed due o he e y hin egion o he p ope y changes
- by he e y na u e o 2D ma e ials and by shielding e ec s in conduc i e
oxides [114]. Simila cha ge doping can also be used o adjus he e ac i e
3.1. OVERVIEW AND STATE OF THE ART 29
index o mul iple quan um well sys ems, in his case he p oblem is opposi e:
The hickness can be inc eased by adding mo e laye s, bu he e ac i e index
change is limi ed [105].
Mo ing away om e ac i e index modula ion and ins ead looking a how
he geome y can be al e ed, one e y emp ing idea is o place he me asu -
ace s uc u es on a s e chable ma e ial, he eby making he pe iod di ec ly
unable [118,119]. Cu en demons a ions a e limi ed in he sense ha i is
ha d o imagine a obus , compac sys em o eliable and high speed ope a-
ion. None heless, hey enable d as ically di e en uning oppo uni ies han
o he me hods, like comple e es uc u ing o he geome y [120] and could be
e y ele an o applica ions such as senso s o moni o ing s uc u e de o -
ma ions [62]. A di e en de o ma ion echnique which is easie o implemen
con ol sys ems o is he mal de o ma ions. The he mal de o ma ion o a
single ma e ial is no much o e nano- o mic oscale dis ances, bu when com-
bining se e al ma e ials wi h di e en he mal expansion coe icien s oge he ,
i is possible o make s uc u es ha mo e conside ably upon empe a u e
changes due o in e nal s ess. Thin bima e ial s uc u es can hus be used
o change angles [121,122] o la e al posi ions [123] o s uc u es ela i e o
each o he . I is also possible o in eg a e hea ing in o he de ice i sel [124],
and o use de o ma ions due o s uc u al changes [125]. Finally, wi hin he
ield o nanoki igami [126] he e is a long lis o mic o- and nanos uc u es
wi h complex de o ma ion pa e ns ha could ha e e y in e es ing op ical
beha io , and ha migh be con olled using MEMS- echniques. The s uc-
u es in [126] o example show a s ong ci cula dich oism ha anishes when
hey a e comp essed using an op ical ibe , while he s uc u es in [127] show a
la ge change in e lec ance when de o med by a pneuma ic p essu e di e ence.
3.1.1 MEMS me asu aces
In MEMS me asu aces, he mechanical mo emen s o MEMS a e used o al-
e he op ical esponse o he me asu ace in some way. The e a e se e al
ad an ages wi h his app oach o designing ac i e me asu aces. Fi s ly, by
no equi ing he op ical change o o igina e in some ma e ial e ec you a e
mo e ee in he choice o ma e ials [128,129]. Fu he mo e, he mechani-
cal mo emen s can be used o make la ge changes in bo h phase and ampli-
ude, and can be in eg a ed wi h high e iciency me asu aces. To d i e hese
mo emen s he e a e ou main ac ua ion echniques, namely elec omagne ic,
piezoelec ic, elec o he mal and elec os a ic [130]. Elec omagne ic based
MEMS gene a e a magne ic ield in o de o in e ac wi h an ex e nal mag-
ne ic ield, achie ing la ge mo emen s a he expense o a less compac sys em.
30 CHAPTER 3. ACTIVE METASURFACES
In some cases his migh be ad an ageous o MEMS only in ended o unc-
ion as mi o s [131], bu o me asu aces hese la ge mo emen s a e o en
no equi ed. Piezoelec ic MEMS a e mo e compac , bu also ha e less ange
and a mo e complica ed ab ica ion p ocedu e, hey a e p esen ed in he nex
sec ion. Elec o he mal MEMS can demons a e la ge mo emen s wi h s ong
o ces, bu ha e limi ed speed and use mo e powe . In e es ingly, hey can
also o e complica ed olding pa e ns, as demons a ed in [124] using a bi-
mo ph s uc u e ha olds in mul iple di ec ions. Meanwhile, elec os a ic
MEMS u ilizes capaci i e o ces be ween elec odes o mo e. They ep esen
he mos common ype o MEMS, as hey a e ela i ely simple o design and
ab ica e, ha e apid esponse imes, low powe consump ion and la ge lexi-
bili y wi h bo h in-plane and ou -o -plane o ces a ailable using comb d i es
and capaci i e pla es. D awbacks include nonlinea i y o he d i ing o ce,
limi ed mo emen ange and equi ing high ol ages o ope a e. In addi ion
o hese ou main ope a ion modes which a e usually ac ua ed elec onically,
i is possible o design MEMS in ended o be op ically ac i a ed, which can be
especially exci ing o in eg a ion wi h me asu aces. Fo example, by making
a MEMS ha abso bs and eac s o one se s o wa eleng hs while moni o ing
i in ano he wa eleng h, se e al g oups ha e demons a ed imaging in mm
and THz wa eleng hs [132–134].
The ”engines” men ioned abo e can be used o gene a e an amazingly
di e se se o mechanical beha io s, bu many o he mos common can be
ca ego ized in o he ollowing:
•Can ile e s a e simple o ealize and can ha e bo h la ge angula and
ou -o -plane mo ion, an in e es ing example is p esen ed in [135], whe e
can ile e s wi h e y la ge mo ion a e used o co e o expose plasmonic
s uc u es which could be used as a a e-ea h-mine al- ee colo display.
•Pis on mo ion, o a mo ion whe e a s uc u e is ansla ed in he ou -
o -plane di ec ion is especially ele an o many op ical applica ions,
and has been used o demons a e high-speed modula ion o abso p ion
equency in bo h he isible [136] and in a ed wa eleng h anges [137].
•Cu ing memb anes a e used in MEMS mi o s o a i ocal e lec i e
lenses [138], and migh be ele an o combining wi h me asu aces.
•In-plane ansla ion, which o example has been used o make a MEMS-
based Al a ez lens [139].
•In-plane o a ion is ano he in e es ing modali y, and has ecen ly been
3.2. PIEZOELECTRIC MEMS 31
used in a MEMS de ice oge he wi h pis on mo ion o p obe he in e -
ac ion be ween laye s o 2D ma e ials [140].
•Tip/ il mo ion is ano he common ype o mo emen , which is use ul o
MEMS mi o scanne s and hus has seen a lo o op imiza ion o la ge
de lec ion angles. By in eg a ing a me asu ace lens on such a mi o ,
a ocusing p o ile can be in eg a ed e icien ly and in he same op ical
plane as he angle de lec ion [141].
Be o e mo ing on o piezoelec ic MEMS we will men ion ha o long
wa eleng hs, MEMS based de ices can be used as indi idual me asu ace pix-
els [142,143]. Howe e , hey a e oo big o do so o he isible and nea
in a ed wa eleng hs, which equi es new de elopmen o mo able s uc u es
o sub-µm sized de ices [128]. A p omising de elopmen in his di ec ion can
be seen in [144], whe e elec on beam li hog aphy and ion beam e ching is
used o make spi al s uc u es wi h a ∼2µm pe iod ha can be de o med
elec os a ically. The s uc u es a e con olled collec i ely bu can be de-
signed wi h indi idual cha ac e is ics, and he ex emely ligh weigh esul s
in a la ge modula ion equency abo e 10 MHz. Figu e 3.8 a he end o his
chap e con ains a compa ison o he e iciency and swi ching speed o some
selec ed me asu ace demons a ions ela ed o he e ec s p esen ed he e, bo h
o MEMS and he o he ac i e me asu ace modali ies.
3.2 Piezoelec ic MEMS
A piezoelec ic ma e ial changes he in e nal pola iza ion when a s ess is ap-
plied, and con e sely becomes s ained when an elec ic ield is applied ac oss
he ma e ial [145]. As such, hese ma e ials a e e y use ul o elec ome-
chanical ansduc ion, and bulk piezoelec ics a e o en used o applica ions
equi ing small bu e y accu a e displacemen s. These bulk piezoelec ic de-
ices usually need high ol ages o ope a e, as he e ec is p opo ional o he
elec ical ield, and a s ong ield ansla es o a la ge ol age i he elec odes
a e a apa . On he o he hand, i he elec odes a e close oge he he oppo-
si e is ue, and hin- ilm piezoelec ics can he e o e be employed o ac ua e
MEMS wi hou he la ge ol ages common in o he applica ions.
One ma e ial o en used o his pu pose is lead zi cona e i ana e, o PZT,
which has conside ably la ge piezoelec ic esponse han mos o he ma e ials
- especially o ce ain mul i-phase polyc ys alline s uc u es [5]. Figu e 3.1
is a schema ic example o a PZT-based MEMS can ile e . An elec ode-PZT-
elec ode s ack is deposi ed on a subs a e and pa s o his subs a e is e ched
32 CHAPTER 3. ACTIVE METASURFACES
away o o m a hin memb ane. When a ol age is applied ac oss he PZT, he
ma e ial - which is clamped o wha is le o he subs a e - expe iences an ou -
o -plane expansion and an in-plane con ac ion. The memb ane is hin in he
ou -o -plane di ec ion, howe e , he elec odes ypically co e a much la ge
a ea and he in-plane s esses he e o e add up o a la ge in-plane s ess on
one side o he memb ane, causing he can ile e o bend. The esul an la ge
mechanical mo emen a low ol ages is one o he main ad an ages wi h hese
hin- ilm based piezoelec ic MEMS, oge he wi h he low leakage cu en
esul ing in ul a-low powe ope a ion on he o de o 10-100 nW depending
on he elec ode su ace a ea and applied ol ages. The disad an age wi h
PZT is ha i has conside able hys e isis, and hus accu a e con ol equi es
some kind o eedback sys em [146].
An example o he ype o ma e ial s ack used o ab ica e he MEMS
mi o s in his wo k is shown in he inse o Figu e 3.1. The s a ing poin is
a silicon on insula o wa e which has a bu ied silicon oxide laye be ween he
silicon handle wa e and he silicon de ice laye . This de ice laye will o m
he passi e pa o he memb ane, while he bu ied oxide is used as a s opping
laye o he deep eac i e ion e ching used o emo e he handle wa e and ee
he memb ane. On op o he de ice laye , a silicon oxide laye is he mally
g own o isola e he elec odes om he silicon while addi ionally ac ing as
a s ess-compensa ion laye . I adjus ed o he co ec hickness his oxide
laye minimizes he inhe en s ess o he inal memb anes, o al e na i ely i
can be used o c ea e some nonze o in insic memb ane cu a u e. On op o
he oxide, he bo om elec ode consis s o a hin i anium adhesion laye , a
pla inum elec ode and a hin laye o LaNiO3. The pla inum o ms a sui able
la su ace o he PZT while also ac ing as a chemical ba ie be ween PZT
and silicon, and he LaNiO3 o ms a seed laye o he PZT deposi ion in
addi ion o imp o ing de ice li e ime by p e en ing he build up o a passi e
laye in he PZT [5]. The ea e , PZT is deposi ed using ei he pulsed lase
deposi ion o chemical solu ion deposi ion, ypically wi h a hickness be ween
1 and 2 µm. Finally, a i anium and wol am adhesion laye and gold op
elec ode a e deposi ed on op, hese do no need a LaNiO3laye as long as
he applied ield is always going om he op o he bo om elec ode. Mo e
de ails on he s ack and he many conside a ions can be ound in [5].
Shown in Figu e 3.2 is a ypical case o how a piezoelec ic MEMS mi o
can be designed. Fou can ile e s wi h piezoelec ic hin ilms suspend a silicon
mass ha has a e lec i e coa ing o ac as a mi o . By applying ol ages o
speci ic a eas o each can ile e , his mi o can hen be mo ed in a pis on-
mo ion along he di ec ion no mal o he mi o su ace, o i can be il ed by
mo ing some can ile e s in one di ec ion and some in he o he . Mo ing he
3.3. COMBINING PIEZOELECTRIC MEMS AND METASURFACES 39
Dynamic wa epla e
Du ing he demons a ion o he swi chable me asu aces men ioned abo e i
became clea ha ce ain aniso opic nanos uc u es gi e a phase upon e-
lec ion ha o one pola iza ion s a e changes d as ically when adjus ing he
ai gap, while o he o hogonal pola iza ion s a e i emains unal e ed. The
e iden idea was hen o use his beha io in o de o ealize a wa epla e
wi h an adjus able phase delay be ween he wo eigenpola iza ion s a es, wi h
he associa ed p oblem being he pola iza ion dependen abso p ion a ce -
ain mi o -me asu ace spacings, when he ligh in e ac s s ongly wi h he
me asu ace s uc u es. A solu ion was ound by choosing a se o nanos uc-
u e dimensions and a la ice pe iodici y ha gi e s ong coupling be ween
esona o s o one pola iza ion s a e and wi h minimal in e ac ion o he o -
hogonal one, he esul ing de ice is p esen ed in [101] ound in Chap e 5.
This s ong coupling gi es nea pe ec e lec ion a he me asu ace laye o
he esonance wa eleng h, and has been explo ed in se e al simila se ings bu
wi h s a ic ca i ies, and wi h a la ge ocus on he ela ed occu ing e ec o
pe ec abso p ion [55,149,150]. Meanwhile, he o hogonal pola iza ion s a e
is ansmi ed h ough he me asu ace and is e lec ed by he mi o . Simply
s a ed, one pola iza ion s a e is e lec ed a he me asu ace and he o he a
he mi o , wi h he sepa a ion dis ance gi ing a ela i e phase shi be ween
he wo, while a mo e accu a e ea men needs o include addi ional e lec-
ions h ough he Fab y-P´e o equa ion. In his way, we we e able o make
a dynamic wa epla e wi h a s able, high and la gely pola iza ion-independen
e lec ion ampli ude wi h ull 2πbi e ingence unabili y.
A ansmissi e e sion would be in e es ing, bu he e a e some undamen-
al challenges. Looking a a simpli ied e sion o he Fab y-P´e o equa ions
o ansmission and e lec ion, he e assuming a ca i y wi h symme ic e lec-
ion and ansmission coe icien s on bo h sides equal o and , and wi h a
p opaga ion phase shi ac oss he ca i y o ϕ, we ha e
o al = + 2 ei2ϕ
1− 2ei2ϕ,
o al = 2eiϕ
1− 2ei2ϕ.
Fo he e lec i e dynamic wa epla e i was ela i ely s aigh o wa d o make
| | ≈ 1 o one pola iza ion and le he o hogonal pola iza ion expe ience
he phase shi om he ca i y, a oiding la ge changes in he ampli ude by
minimizing | 2|. Howe e , in ansmission bo h pola iza ions will necessa ily
expe ience he p opaga ion phase ac oss he ca i y a leas once. I is o cou se
40 CHAPTER 3. ACTIVE METASURFACES
possible o make one pola iza ion s a e in e ac s ongly wi h he ca i y and
he o he no , bu his s ongly a ec s he ansmission ampli ude o he in-
e ac ing pola iza ion s a e, hus educing e iciency - and pe haps e en mo e
p oblema ically in oducing une en e iciencies be ween he di e en pola iza-
ion s a es. A solu ion migh be o no wo k in he Fab y-P´e o egime bu
a he using nea ield coupling, o example using a sys em along he lines
o [43], whe e an a ay o plasmonic an ennas a e in close p oximi y o a me al
laye wi h an a ay o holes. None heless, i is unclea how one would keep a
low e lec ion o all se ings and simul aneously achie e a ull 2πphase ange.
Fab y-P´e o ope a ion
A icle 3 in Chap e 3, [151], compa es he cha ac e is ics o he MEMS-
me asu ace pla o m when ope a ing in he gap su ace plasmon and Fab y-
P´e o ope a ion egimes. The e iciencies a he design wa eleng h a e shown
o be e y simila , wi h only sligh deg ada ion when going o la ge ca -
i y leng hs. Op imal s uc u e sizes o he pola iza ion independen blazed
g a ing which was in es iga ed a e also nea ly iden ical, wi hin no mal ab i-
ca ion unce ain ies a op ical wa eleng hs. The main di e ence pe ains o
he bandwid h, which ge s educed a la ge ca i y leng hs due o he wa e-
leng h dependence o he Fab y-P´e o esonances. Compa ing he nea ield
ope a ion wi h ope a ing a he i s Fab y-P´e o esonance, he bandwid h
was ound o all om a ound 20 % o 7 %.
Tunable me alens double
U ilizing he same ype o MEMS, bu in a di e en con igu a ion, he de ice
in [152] consis s o wo dielec ic me asu ace lenses ope a ing as a ansmissi e
double , in a way simila o wha is done in [153]. One me asu ace lens is
moun ed in a MEMS ame, basically a MEMS mi o wi h he mi o pa
emo ed, see Figu e 3.5. This allows he dis ance be ween he wo lenses
o be al e ed, he eby changing he o al ocal leng h o he sys em. The
wo me asu ace lenses a e iden ical, and consis o silicon pilla s placed in a
geome ic phase pa e n. Thus, he ci cula ly pola ized ligh inciden on he
i s lens is c oss-pola ized be o e hi ing he nex lens. As he me asu ace is
based on he geome ic phase, his c oss-pola ized ligh would ha e expe ienced
a lipped phase bu o he ac ha his second lens is lipped, e e sing he
o a ion angles o he nanos uc u es, as seen om he side o he incoming
ligh . In his way a MEMS displacemen o 7 µm is con e ed o a ocal lengh
shi o 250 µm.
3.3. COMBINING PIEZOELECTRIC MEMS AND METASURFACES 41
Figu e 3.5: One o wo me asu ace lenses combined o o m a double . The
ape u e o he squa e me alens seen in he cen e o he pho o is 1.5 mm.
The o he me alens (no shown) was kep s a ic, while he one shown in he
pic u e is placed in a MEMS ame ha can be ansla ed by abou 7 µm o
change he sepa a ion o he wo lenses - hus adjus ing he o al ocal leng h
o he double . The lowe le inse is a SEM image o he silicon pilla s used
o implemen he geome ic phase lens. The de ice was documen ed in [152].
42 CHAPTER 3. ACTIVE METASURFACES
Dynamic linea pola ize
Du ing he discussion o he dynamic wa epla e abo e, i was men ioned ha
s ong coupling was in es iga ed in simila sys ems in connec ion wi h pe ec
abso p ion. In [154] his is used o c ea e a a iable linea pola ize , whe e he
e lec ion o one linea pola iza ion s a e is cons an ly a ound 98 %, while o
he o hogonal s a e i is a ied om 95 % o 7 %. The ex inc ion a io is hus
adjus ed om 1 o 13, wi h he simula ions sugges ing an ex inc ion a io abo e
60 could be possible wi h mo e accu a e ab ica ion and moun ing. Looking
a o he in es iga ions, [149] was o example able o achie e 3 % e lec ance,
which would co espond o an ex inc ion a io unabili y o 32 in ou case.
La ge ci cula dich oism
The wa epla e and linea pola ize p esen ed abo e ha e linea pola iza ion
s a es as eigens a es o he me asu ace, meaning ha i one o hese s a es
a e inciden he e lec ed s a e is he same bu wi h an al e ed phase and/o
ampli ude. In [53] an asymme ic uni cell consis ing o ou sepa a e plas-
monic esona o s is used o make a e lec i e de ice ha can unc ion as a
s anda d mi o when he MEMS is in a speci ic posi ion. Meanwhile, a a
second MEMS posi ion he e is a la ge ci cula dich oism whe e one ci cula
pola iza ion s a e is ully abso bed while he o hogonal s a e is e lec ed and
c oss-pola ized, as compa ed wi h he e lec ion om a mi o . The a icle
explo es he e ec in he amewo k o excep ional poin s [155], whe e he e is
a simul aneous degene acy in bo h he eigens a es and eigen alues.
Mode swi ching lase
In Sec ion 2.1.3 he use o s a ic me asu aces o mode con ol in lase ca i-
ies was men ioned. This is demons a ed wi h an ac i e me asu ace in [156],
whe e a e lec i e MEMS-me asu ace is used o make a lase in which he
emi ed ligh can be swi ched be ween a Gaussian mode and a mode wi h
o bi al angula momen um. The me asu ace uses a simila nanos uc u e ge-
ome y as o he dynamic wa epla e, bu wi h dimensions scaled o unc ion
a a wa eleng h o 1030 nm, hese s uc u es a e hen o a ed o impa a
geome ic phase acco ding o he desi ed deg ee o o bi al angula momen-
um. Thus, when mo ing he MEMS mi o he beha io o each indi idual
me asu ace cell can be swi ched be ween ha o a hal -wa epla e and a mi -
o , gi ing he in ended e ec . Ampli ica ion is done using a pola iza ion
main aining ibe , which is connec ed o a ee-space ca i y inco po a ing he
3.3. COMBINING PIEZOELECTRIC MEMS AND METASURFACES 43
MEMS-me asu ace. By using he e lec i e de ice a a sligh ly oblique inci-
dence (7◦) i is possible o ha e a Gaussian mode o coupling in o he ibe on
one side o he MEMS-me asu ace, while keeping a o bi al angula momen-
um ca ying s a e on he o he side, whe e an ou pu couple emi s he lase
ou pu .
Bilaye me asu aces
Wi h a single me asu ace laye combined wi h a MEMS mi o i is possible o
swi ch be ween mi o beha io and ha which is buil in o he me asu ace.
Fo some cases, such as he dynamic wa epla e, his migh en ail se e al use ul
unc ionali ies o e en a con inuous ange. Howe e , you only ha e indepen-
den con ol o e one designed beha io . Fo example, when using a blazed
g a ing me asu ace, adjus ing he ca i y spacing will only al e he di ac-
ion o de e iciencies. By adding a second me asu ace laye , his limi a ion
is emo ed and i is possible o inco po a e wo independen unc ionali ies,
his is done in [157] - which is included in Chap e 5. The basic p inciple is
depic ed in Figu e 3.6, whe e i is explained how a me asu ace can be placed
a he node o a s anding wa e pa e n, o med when a monoch oma ic beam
is no mally inciden on he me asu ace-mi o sys em. In his case he sys-
em ac s as a s anda d mi o , and he me asu ace can e en be in isible in a
mic oscope image. When mo ing he mi o , he me asu ace is no longe a
he node and he designed unc ionali y is swi ched on.
Going back o ha ing he me asu ace a he node, i is also possible o add
a second me asu ace which will in e ac wi h he ield, esul ing in wha e e
beha io his second me asu ace is designed o . Changing he mi o posi-
ion now, howe e , will cause he ield o in e ac wi h bo h o he laye s, e en
when he sepa a ion be ween he wo is se equal o λ/4, hough his educes
c oss- alk. To accoun o hese in e ac ions, including all he mul iple e lec-
ions and ansmissions, a ans e ma ix me hod is used [150]. The a icle
documen s how his is done success ully o se e al di e en combina ions o
me asu ace unc ionali ies, wi h a key akaway being ha he ealized quali y
is be e o he unc ionali y only dependen on one o he me asu ace laye s.
Also no able is he swi ching speed achie ed, using a small MEMS mi o ise
and all imes o a ound 5 µs a e measu ed.
Me asu ace pola ime e
Con a y o he i le o his hesis, he las de elopmen is no an ac i e me a-
su ace bu a he an in eg a ed sys em based on a p inciple i s desc ibed
44 CHAPTER 3. ACTIVE METASURFACES
Figu e 3.6: A no mal incidence, he e lec ion o a monoch oma ic wa e a
a me allic in e ace causes a s anding wa e pa e n be ween he incoming and
e lec ed wa e. Placing a me asu ace in he ex ema o he s anding wa e
pa e n as in awill gi e e y di e en esul s han placing in he minima b,
whe e he ield in p ac ice does no see he me asu ace. In ha way, cis
p ac ically iden ical o he case in a, bu i he ca i y leng h is al e ed he
ield will in e ac wi h bo h me asu aces. An example o his using a single
laye me asu ace is shown in he wo images: dA one ca i y leng h he
me asu aces ( wo la ge and six small ci cles) a e isible; dwhen he MEMS is
used o change ca i y leng h he me asu aces a e no longe isible, al hough
he alignmen ma ke (c oss) is s ill isible on he bo om le due o a sligh
il o he mi o . A ho ough e iew on such ca i y e ec s can be ound
in [55].
3.3. COMBINING PIEZOELECTRIC MEMS AND METASURFACES 45
Figu e 3.7: Showcasing o a me asu ace pola ime e a SPIE Pho onics Eu-
ope 2024. The me asu ace pola ime e is documen ed in [158].
in [6], whe e a me asu ace is used o spli an incoming beam in o six di ac-
ion spo s, he in ensi ies o which a e used o de e mine he pola iza ion s a e
o he incoming beam. The sys em, consis ing o a me asu ace, beam-spli e ,
came a and alignmen componen s is p esen ed in [158], o be ound in Chap-
e 5. The design wa eleng h is chosen o be 640 nm, o make i sui able
o in eg a ion in a se up o digi al pola ime ic his opa hology [159]. This
echnique analyzes pola iza ion changes om ligh sca e ed by issue, wi h
he goal o speeding up and imp o ing medical diagnos ic p ocedu es.
A 640 nm he abso p ion o gold s a s being no iceable, and educed
e iciency is clea ly shown in he simula ions o sho e wa eleng hs. S ill, a
he nominal wa eleng h he sys em wo ks well and is benchma ked agains a
comme cially a ailable pola ime e wi h good esul s, accu a ely de e mining
he S okes-componen s and he deg ee o pola iza ion wi hin 2 %. Howe e ,
o e icien use in he digi al his opa hology se up wi h eal issue samples,
i was de e mined ha he me asu ace pola ime e lacks su icien dynamic
ange, and i is he e o e sugges ed o eplace he image senso wi h indi idual
diodes. In addi ion o imp o ing dynamic ange, his would g ea ly imp o e
acquisi ion speed - especially compa ed o adi ional pola ime e s ha employ
a physically o a ing qua e -wa epla e.
46 CHAPTER 3. ACTIVE METASURFACES
3.4 MEMS me asu aces, a compa ison
Table 3.1 is an o e iew o some key pa ame e s o he piezoelec ic MEMS
me asu aces discussed in he p e ious sec ion. Rega ding he e iciencies hey
a e all signi ican , wi h he excep ion o he a i ocal lens double , bu his
was in no pa due o he MEMS as bo h he silicon chips lacked an i e lec ion
coa ing on he backside, and he c oss-pola iza ion e iciency o he s uc u es
in bo h lenses had oom o imp o emen . The o he e iciencies a y qui e
a bi , om 30 % o he bilaye me asu aces o 95 % o he dynamic linea
pola ize s, e lec ing he la ge ange o di e en e ec s demons a ed. Simila
esul s o be e a e expec ed i mo ing o longe wa eleng hs, whe e one has
he combined ac o s o less (unwan ed) plasmonic losses, as well as highe ol-
e ance on wha cons i u es a sho ca i y leng h o imp o ed bandwid h. Fo
sho e wa eleng hs in he isible, one quickly uns in o la ge esis i e losses
in gold, al hough he me asu ace pola ime e demons a es i is s ill a iable
op ion o ed ligh . Using aluminium, sil e o conduc i e oxides as discussed
in Sec ion 2.1.2 could be an op ion o hese wa eleng hs. Al e na i ely one
could also swi ch o dielec ics, he me asu aces in [30] o example consis o
dielec ic pilla s in close p oximi y o a me allic mi o , and a e used o eely
design bo h he e lec ed phase and ch oma ic dispe sion o ligh . The s udy
is o wa eleng hs be ween 1440 nm and 1590 nm and uses high dispe sion
silicon me a-a oms, bu he design could be scalable o he isible, especially
i conside ing swi ching o silicon ich ni ide - whe e one can une he depo-
si ion p ocess pa ame e s o ade e ac i e index con as and dispe sion o
less abso p ion [40].
When i comes o swi ching speed, he de ices a e all based on simila
MEMS and ha e oughly he same swi ching speeds o a oud 100-400 µs wi h
wo excep ions. Fo he linea pola ize we suspec he e was some de ec in
he MEMS causing sligh ly slowe esponse, meanwhile, o he bilaye me a-
su aces a smalle mi o (500 µm diame e ) wi h sho e memb ane was used,
esul ing in a swi ching speed o only 5 µs. No e ha his MEMS could be
used wi h any o he o he concep s o achie e he same a e. Excep o
he a i ocal double me alenses made using nanoimp in li hog aphy, all he
me asu aces ha e been made using e-beam li hog aphy wi h a o al diame e
o a ound 100 µm. Fo demons a ion pu poses he MEMS mi o diame-
e has he e o e no been a limi ing ac o , bu his migh be he case o
eal-wo ld applica ions, whe e one can end up wi h ade-o conside a ions
be ween swi ching a e and he ee ape u e size. Rega ding he limi s o his
pla o m, we ha e wo ked wi h diame e s om 100 µm o 5 mm, and wi h
mm-sized ape u es he swi ching speed o hese mi o s lies in he ange o
3.4. MEMS METASURFACES, A COMPARISON 47
Table 3.1: An o e iew o some main cha ac e is ics o he de eloped MEMS
and me asu ace componen s. The componen abb e ia ions e e o he p e-
ious sub-sec ions and s and o : gap su ace plasmon (GSP), dynamic wa e-
pla e, Fab y-P´e o based me asu ace, unable lens double , linea pola ize
(LP), ci cula dich oism (CD), mode swi ching lase , bilaye me asu ace and
me asu ace pola ime e . The ca ego ies a e e iciency, swi ching ime ( s), de-
sign wa eleng h (λ), bandwid h (BW) and minimum ca i y leng h (min(Ta)).
a100 nm bandwid h is o he unc ionali y whe e only one me asu ace con-
ibu es, he o he unc ionali y has a sho e bandwid h.
bBandwid h is s ongly skewed owa ds longe wa eleng h han he design
wa eleng h, due o s ong abso p ion o gold in he isible equencies.
E . % s[ms] λ[nm] BW [nm] min(Ta) Re .
GSP 56 0.4 800 160 50 nm [148]
Wa epla e 75 0.4 800 100 2.2 µm [101]
FP 40 - 800 100 - [151]
Double low ∼0.1 1550 - - [152]
LP 95 2 831 <10 3.4 µm [154]
CD 50 0.2 810 <10 1.5 µm [53]
Mode swi ching 80 0.1 1030 - 465 nm [156]
Bilaye 30 0.005 800 100a1.7 µm [157]
Pola ime e 50 S a ic 640 >100b- [158]
48 CHAPTER 3. ACTIVE METASURFACES
10-20 kHz. The e is oom o some imp o emen wi h adjus men s such as
hinning he handle wa e sec ion o he mi o mass, and ading mo emen
ange o highe esonance equency by making a s i e memb ane. In he
limi o hese adjus men s he MEMS will s a o esemble a PMUT (piezo-
elec ic mic omachined ul asound anduce ), and om a ecen e iew a i-
cle [160] i seems likely ha ele an designs will be es ic ed o he 100 kHz
ange, al hough he combina ion o la ge ape u e and sho de lec ion is ou -
side he scope o mains eam PMUT esea ch. No e howe e , ha i going
o smalle ape u es PMUTs a e able o go high up in equency while gi ing
su icien mo emen o swi ching o me asu aces. To each MHz equencies
wi h MEMS using la ge ape u es i p obably necessa y wi h dedica ed de-
signs such as p esen ed in [136], whe e he me asu ace unc ionali y is buil
in o a suspended silicon s uc u e ha can be mo ed elec os a ically. In o de
o each high modula ion equencies his s uc u e is also limi ed o ha ing
a small ape u e, bu he elec os a ic ac ua ion gi es a de ice which can be
mo e easily pa e ned in o an a ay wi hou oo much non- unc ional a ea.
Figu e 3.8 compa es a selec ion o ac i e me asu aces in he e iciency-
swi ching speed space, including he MEMS me asu aces om Table 3.1, while
highligh ing wha kind o basic p inciple is used o modula ion. In his iew
he ad an ages o he demons a ed pla o m a e clea , i can be used o ealize
ac i e me asu aces wi h high e iciency and wi h apid esponse imes. I is
no sui able o ul a- as modula ion, bu unlike some o he as e e ec s he
MEMS-based app oach p o ides ull phase and ampli ude unabili y, and wi h
good bandwid h o many unc ionali ies. Compa ed o o he MEMS me asu -
aces he me hod o moun ing a me asu ace oge he wi h a gene al MEMS
is qui e a ac i e, as i allows he ela i ely complex MEMS p oduc ion o be
decoupled om he speci ic applica ion he me asu ace is designed o . In his
way one p oduc ion un o MEMS can be used o a ious applica ions, which
has been especially use ul o p o o yping - his selling poin migh be mo e
impo an o piezoelec ic han o elec os a ic MEMS, due o he mo e com-
plica ed ab ica ion. Disad an ages, on he o he hand, include he collec i e
modula ion, o he lack o pixel-by-pixel con ol which applies o mos o he
concep s oo. By using a e lec i e mi o on he MEMS, he demons a ed
de ices a e also limi ed o wo king in e lec ion, o applica ions whe e his is
app op ia e, howe e , i is an ad an age since i imp o es e iciency. Losses in
he isible equencies a e also a challenge, bu should be possible o add ess
by using o he ma e ials. Fu he mo e, due o hys e isis in he piezoelec ic
ma e ial, some kind o eedback mechanism is necessa y o accu a e con ol,
his adds a laye o complexi y bu can be done in se e al ways, o example
wi h piezo esis i e elemen s, capaci i e measu emen s o op ical ead ou [146].
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
2 o 11
ul a la MEMS mi o se ing as a mo eable back e lec o (Fig.1A).
OMSs and MEMS mi o s a e designed and ab ica ed in sepa a e
p ocessing pa hs and hen combined, ensu ing he eby he design
eedom on bo h sides and educing he ab ica ion complexi y. The
choice o he piezoelec ic MEMS o be combined wi h he GSP-
based OMS is dic a ed by speci ic ad an ages o he o me , including
con inuous ou -o -plane ac ua ion capabili y and low ol age/powe
ope a ion (53), which enable he de elopmen o con inuously unable/
econ igu able MEMS-OMS componen s wi h ul acompac sizes
and low powe consump ion.
Wi h his pla o m, we expe imen ally demons a e dynamic
pola iza ion-independen beam s ee ing (Fig.1B) and e lec i e 2D
ocusing (Fig.1C). By elec ically ac ua ing he MEMS mi o and
hus modula ing he MEMS-OMS dis ance, pola iza ion-independen
dynamic esponses wi h la ge modula ion e iciencies a e demon-
s a ed. Speci ically, when ope a ing a a wa eleng h o 800 nm, he
beam s ee ing e iciency (in he +1s di ac ion o de ) eaches 40 and
46% o he espec i e ans e se magne ic (TM) and ans e se elec-
ic (TE) pola iza ions (elec ic ield pa allel/pe pendicula o he
e lec ion plane, espec i ely), whe e 76 and 78% a e expec ed om
simula ions, while he beam ocusing e iciency eaches 56 and 53%
(64 and 66% expec ed om simula ions). Fu he mo e, he dynamic
esponse o he in es iga ed MEMS-OMSs is cha ac e ized wi h he
espec i e ise/ all imes o ~0.4/0.3 ms, cha ac e is ics ha can be
u he imp o ed by using MEMS mi o s op imized o bandwid h
in he megahe z ange. Fo example, by using MEMS ac ua ed
memb anes o ensu e ~30MHz o swi ching speeds (54–56).
RESULTS
Ope a ional p inciple
Simila o he con en ional GSP-based OMSs (6–8,57), he p oposed
MEMS-OMS con igu a ion ep esen s a me al-insula o -me al (MIM)
s uc u e composed o a bo om hick gold laye a op a silicon sub-
s a e (MEMS mi o ), an ai space , and a op laye wi h 2D a ays
o gold nanob icks on a glass subs a e (OMS s uc u e). The ai
space gap a can be inely adjus ed by ac ua ing he MEMS mi o
(Fig.2A). When he ai gap is small ( a<200 nm), he op ical e-
sponses o OMS uni cells a e de e mined by he GSP exci a ion and
esonance in he MIM con igu a ion (57,58) and hus by nanob ick
dimensions (8,57). To p og ess u he owa d he design o dy-
namically con olled MEMS-OMSs, se e al geome ical OMS pa-
ame e s mus be de e mined. Fi s , we se he ope a ing wa eleng h
a 800nm and choose he OMS uni cell size o 250nm ha should
be subs an ially smalle han he ope a ing wa eleng h (8,57).
Assuming he smalles achie able ai gap is be ween 20 and 50 nm,
he nanob ick hickness m is hen op imized o achie e a wide phase
co e age wi h la ge e lec ion ampli udes, esul ing in he choice o
m=50nm ( ig. S1). The nanob ick la e al dimensions, side leng hs,
a e chosen o be equal o ensu e he pola iza ion-independen op i-
cal esponse. Analysis o he complex e lec ion coe icien s o he
OMS conduc ed o inc eased ai gaps e eals ha he phase g adien
o di e en nanob ick side leng hs p og essi ely dec eases, wi h
he e lec ion phase and ampli ude becoming independen on he
nanob ick leng h a an ai gap o a=350nm (Fig.2,BandC). This
d as ic ans o ma ion in he op ical esponse is ela ed o s ong
dependencies o he GSP exci a ion (a no mal incidence) and GSP
e lec ion a nanob ick e mina ions on he ai gap: Bo h dec ease
apidly o inc eased ai gaps, he eby a enua ing and e en ually
elimina ing he GSP esonance. The obse ed ans o ma ion o he
e lec ion phase esponse (Fig.2C) implies a simple and s aigh o wa d
app oach o ealize dynamically con olled MEMS-OMSs: Fo a gi en
smalles ai gap ( o example, 20 nm), one can design any concei -
able GSP-based OMS (59), whose unc ionali y can hen be swi ched
on and o by mo ing he MEMS mi o . He ea e , we demons a e
his app oach by ealizing dynamically con olled pola iza ion-
independen beam s ee ing and e lec i e 2D ocusing.
Pola iza ion-independen dynamic beam s ee ing: Design
The MEMS-OMS design o ealizing dynamically con olled
pola iza ion-independen beam s ee ing equi es he choice o he
numbe N o uni cells in he OMS supe cell ha , in u n, de e -
mines he s ee ing angle o he gi en uni cell size =250 nm,
e ac i e index o silica glass n=1.46, and ligh wa eleng h =800 nm:
sin =/nN (6,8,57). Bea ing in mind expe imen al condi ions,
we chose an OMS supe cell consis ing o 12 cells so ha he s ee ing
angle is =10.5° in glass (co esponding o 15.5° in ai ), acili a ing
he cha ac e iza ion o well-sepa a ed 0 h/±1s di ac ion o de s
Fig. 1. 2D wa e on shaping wi h he MEMS-OMS. (A) Schema ic o mi o -like ligh e lec ion by he MEMS-OMS be o e he ac ua ion, i.e., wi h he ini ial gap
o ~350 nm be ween he OMS nanob ick a ays and MEMS mi o . Inciden ligh is specula ly e lec ed by he MEMS-OMS ega dless he OMS design. (B and C) Schema ic
o demons a ed unc ionali ies, (B) anomalous e lec ion and (C) ocusing (depending on he OMS design), ac i a ed by b inging he MEMS mi o close o he OMS
su ace, i.e., by dec easing he ai gap o ~20 nm.
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
3 o 11
wi h a 20×/0.42 objec i e. Following he app oach desc ibed abo e,
he phase esponse calcula ed wi h he ai gap a=20nm o di e en
nanob ick leng hs is used o selec he 12 nanob icks (Fig.2C,
ed ci cles) and a ange hem in o an a ay along he x di ec ion
(Fig.2,DandE) o mimic he e lec ion coe icien o an ideal
blazed g a ing: (x)=Aexp(i2x/sc) (6,8,57), whe e A≤1 is he
e lec ion ampli ude, and sc=12 is he g a ing (supe cell) pe iod.
The a ailable phase ange a a=20nm is sligh ly mo e han 270°
( he ed dashed line in Fig.2C), implying ha i is impossible o
design a supe cell wi h 12 di e en uni cells ensu ing a cons an
phase g adien ( he la e equi es he phase ange o 11×30°=330°).
One possible app oach o deal wi h his p oblem is o inc ease he
phase (disc e iza ion) s eps o 90° ( ig. S2, A o E), so ha he equi ed
phase ange would dec ease o 3×90°=270°, esul ing in he pos-
sibili y o compose he supe cell om duplica ed (sc=4×2=8)
o iplica ed (sc=4×3=12) cells (8). Ou simula ions sugges ed
ano he app oach, in which wo uni cells we e le ou emp y, i.e.,
wi hou nanob icks, while he o he 10 nanob icks co e he a ail-
able phase ange o 270°, hus ensu ing be e sampling o he phase
p o ile and imp o ing he e iciency o di ac ion o he desi ed +1s
o de ( ig. S2, F o L). No e ha , in he absence o abso p ion, one
migh op o ano he app oach, such as doubling only he cells wi h
ex eme (minimum and maximum) phase esponses (60).
The e lec ed elec ic ield (x/y componen s) calcula ed o hus
designed MEMS-OMS unde he TM/TE inciden ligh a 800-nm
wa eleng h wi h a=20nm mani es s smoo h wa e on s a eling
in he di ec ion o he +1s di ac ion o de (Fig.2F and ig. S2J, le ).
Fo inc eased ai gaps, he phase g adien s p oduced by he supe cell
nanob icks p og essi ely dec ease as expec ed (Fig.2,BandC), wi h
he phase g adien becoming ze o and he e lec ed ield e u ning
o he specula e lec ion a an ai gap o a=350nm (Fig.2G and
ig. S2J, igh ). He e, we ema k ha ou simula ions p esen ed
Fig. 2. Pola iza ion-independen dynamic beam s ee ing: Design. (A) Schema ic o he OMS uni cell including he ai gap and gold mi o . (B) The complex e lec ion
coe icien calcula ed as a unc ion o he nanob ick side leng h Lx and ai gap a wi h o he pa ame e s being as ollows: = 800 nm, m = 50 nm, = 250 nm, and Ly = Lx.
Colo a ion is ela ed o he e lec ion ampli ude, while he magen a lines ep esen cons an e lec ion phase con ou s. (C) Re lec ion phase (dashed lines) and ampli ude
(solid lines) dependencies on he nanob ick leng h Lx o wo ex eme ai gaps: a = 20 nm ( ed) and 350 nm (blue). Ci cles ep esen he nanob ick sizes selec ed o he
OMS supe cell designed o dynamic beam s ee ing. (D) Top iew and (E) c oss sec ion o he designed MEMS-OMS supe cell. (F and G) Dis ibu ions o he e lec ed TM
elec ic ield (x componen ) a 800-nm wa eleng h o ai gaps o a = 20 and 350 nm, espec i ely. (H) Di ac ion e iciencies o di e en o de s (|m| ≤ 1) calcula ed as a
unc ion o he ai gap a o TM/TE inciden ligh wi h 800-nm wa eleng h. (I) Di ac ion e iciencies o di e en o de s (|m| ≤ 1) calcula ed a he ai gap a = 20 nm as a
unc ion o he wa eleng h o TM/TE inciden ligh .
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
4 o 11
he ea e a e conce ned wi h he ai gaps limi ed by 350nm since,
o la ge ai gaps, a MEMS-OMS would unc ion in a comple ely
di e en egime de e mined by mul iple and pe iodic posi ions o
Fab y-Pé o esonances (see Discussion). The associa ed dec ease in
he +1s o de di ac ion e iciency and inc ease in he 0 h o de
one as a unc ion o he ai gap, a e p ac ically linea , p omising la ge
modula ion e iciencies a ailable wi h he ac ua ed MEMS-OMS
(Fig.2H). Thus, +1s /0 h-o de di ac ion e iciencies a e expec ed
o change om ~77/0 o 0/96% ( o bo h TM and TE pola iza ions)
when changing he ai gap om 20 o 350nm. Redis ibu ions o he
powe be ween di ac ed o de s o g adually a ying ai gaps a e
in e connec ed wi h he co esponding modi ica ions in he e lec ed
ields, unde going g adual ansi ion ( ig. S3) be ween hose p ima ily
di ac ed (a a=20 nm) and hose p ima ily e lec ed (a a=350 nm).
The designed MEMS-OMS is expec ed o exhibi he b oadband
ope a ion simila o ha known o con en ional GSP-based OMSs
(7,8,57). We no e ha he MEMS-OMS pe o mance a la ge ai
gaps is equi alen o ha o a mi o , wi h he alue o a sui ably
la ge ai gap being p opo ional o he ope a ing wa eleng h (see he
conside a ion o he Fab y-Pé o –based ope a ion in Discussion).
Wi h his ca ea in mind, he MEMS-OMS o e all pe o mance is
de e mined by ha a he smalle ai gap o a=20 nm, sugges ing a
1-dB bandwid h o ~150nm nea he ope a ing wa eleng h o
800nm (Fig.2I). No e ha , while he e lec ed ield dis ibu ion o
he ai gap o 20nm (Fig.2F and ig. S2J, le ) is no ideal o a numbe
o easons: insu icien phase ange, unequal ampli ude e lec ion
coe icien s, e c. (7,8,57), he pe o mance o he MEMS-OMS a
he design wa eleng h o 800nm is p ac ically ideal wi h only he
+1s di ac ion o de being nonze o (Fig.2I and ig. S2, H and I),
i.e., his nonideal wa e on o ma ion is o no p ac ical impo ance
o he de ice ope a ion. As a inal commen , i should be men ioned
ha , gi en he possibili y o small ai gap adjus men s a ound he
designed ai gap o a=20 nm, he di ac ion e iciencies o di e en
wa eleng hs could be enhanced, hus imp o ing he e ec i e band-
wid h o he MEMS-OMS de ice ( ig. S2, K and L).
Pola iza ion-independen dynamic beam s ee ing:
Cha ac e iza ion
The MEMS-OMS o pola iza ion-independen dynamic beam s ee -
ing designed abo e (Fig.2) was assembled om a sepa a ely
ab ica ed OMS, an ul a la MEMS mi o (53), and a p in ed ci cui
boa d (Fig.3A; o de ails, see Ma e ials and Me hods along wi h
ig. S4, A o D). The possibili y o ab ica ing he MEMS mi o and
OMS sepa a ely simpli ies he design and ab ica ion p ocesses, o
example, by allowing he wo componen s o be p oduced in sepa-
a e p ocessing lines wi h di e en minimum linewid h capabili ies.
The ab ica ed MEMS mi o and OMS we e cha ac e ized indi idually
using an op ical mic oscope and scanning elec on mic oscope (SEM)
(Fig.3,BandC). When joining he MEMS mi o and OMS, i is
impo an o a oid any pa icles ha can obs uc he mi o om
ge ing close enough o he OMS. Because he mi o (i.e., 3mm in
diame e ) was much la ge han he OMS (i.e., 30 m by 30 m in
size), he OMS was ab ica ed on op o a 10-m-high pedes al, he
idea being ha any pa icles smalle han 10m ou side he pedes al
will no p e en he OMS and MEMS mi o om coming in o con-
ac . This pedes al did no a ec he ab ica ion o he nanob icks,
ea u ing o e all consis ency wi h he design apa om sligh ly
ounded co ne s and mino size de ia ions (Fig.3C) ha a e no
expec ed o p oduce no iceable de e io a ion in he OMS pe o m-
ance (8). A e assembling he MEMS-OMS, he MEMS-OMS
sepa a ion was es ima ed using whi e ligh in e e ome y (Zygo
NewView 6000) o be ~2m ( ig. S4E), which is well wi hin
he ~6-m-la ge mo ing ange o he MEMS mi o (see Ma e ials
and Me hods along wi h ig. S4F). Following ha , we es ima ed he
smalles achie able sepa a ion be ween he MEMS mi o and OMS
subs a e su ace (c ucial o e icien modula ion) by using a
mul iwa eleng h in e e ome y ( ig. S5). We ound by ac ua ing he
MEMS mi o ha , o se e al assemblies, his gap ( m+ a) can be
as small as ~100nm ( ig. S5), co esponding o a~50 nm, and hese
samples we e hen selec ed o u he op ical cha ac e iza ions.
To cha ac e ize he MEMS-OMS pe o mance, we used a
wa eleng h- unable (~700 o 1000 nm) lase wi h he co esponding
op ical, pola iza ion, and imaging componen s (see Ma e ials and
Me hods along wi h ig. S6). The MEMS mi o is elec ically ac u-
a ed o modula e he op ical esponse o he MEMS-OMS obse ed
isually in bo h di ec objec (OMS su ace) and Fou ie image
planes (Fig.4A). In he di ec objec images, his e ec o powe
edis ibu ion is seen in he appea ance (a nonze o ac ua ion ol ages)
o well-p onounced in e e ence inges o med due o he in e e ence
be ween he esidual specula e lec ion and he +1s -o de di ac ed
Fig. 3. MEMS-OMS assembly. (A) Typical pho o o he MEMS-OMS assembly consis ing o he OMS pa e ned on a glass subs a e, an ul a la hin- ilm MEMS mi o , and
a p in ed ci cui boa d (PCB) o elec ical connec ion. (B) Op ical mic oscopy and (C) SEM images o he OMS ep esen ing he 30 m by 30 m and 250-nm-pe iod a ay
o di e en ly sized gold nanob icks designed o dynamic beam s ee ing, ab ica ed a op a 10-m-high pedes al on he glass subs a e, and used in he MEMS-OMS
assembly. Pho o c edi : Chao Meng, Uni e si y o Sou he n Denma k.
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
5 o 11
beam. Fo bo h pola iza ions, he edis ibu ion o adia ion powe
be ween he 0 h and +1s di ac ion o de s a e well p onounced,
eaching he maximum con as a 3.75 V wi h he di ac ion
e iciencies o 40/46% o he espec i e TM/TE pola iza ions
(Fig.4,AandB). The expe imen ally ob ained di ac ion e iciencies
(Fig.4B) a e no iceably smalle han hose expec ed om he simu-
la ions (Fig.2H), disc epancies ha a e somewha expec ed and
a ibu ed o addi ional abso p ion in gold nanob icks because o
su ace sca e ing and g ain bounda y e ec s as well as inc eased
damping associa ed wi h a nanome e - hin i anium adhesion laye
be ween gold-glass in e aces (8). No e ha he e is also a mino
di e ence o be expec ed because o di e en media conside ed
when de e mining he heo e ical and expe imen al e iciencies (see
Ma e ials and Me hods). The high-con as dynamic beam s ee ing,
induced by ac ua ing he MEMS mi o wi h he al e na ing ol ages
o 0 and 3.75 V a a slow swi ching speed, is clea ly seen in he mo ie
cap u ed by he cha ge-coupled de ice (CCD) came a (mo ie S1).
The MEMS-OMS ope a ion is ound o be pola iza ion independen
and b oadband, exhibi ing he 1-dB bandwid h o ~150nm (Fig.4C).
By ac ua ing he MEMS mi o wi h a pe iodic ec angle signal and
de ec ing he spa ially sepa a ed 0 h/+1s o de o di ac ion ields,
one obse es ela i ely as swi ching wi h he ise/ all imes
o ~0.4/0.3 ms, espec i ely (Fig.4D). The esponse speed is ela ed
o he in insic oscilla ion equency o he MEMS mi o , hus being
dependen on he MEMS design pa ame e s such as geome y, weigh ,
s i ness, and so on (53–56). No e ha he s anda d hin- ilm MEMS
mi o used is a he la ge (~3mm in diame e ; Fig.3A), wi h i s
su ace a ea o de s o magni ude la ge han ha o he OMS a ea
(~30 m by 30 m in size; Fig.3B), conside ably slowing down he
dynamic esponse. Bea ing in mind he possibili y o op imizing he
MEMS mi o o as swi ching speeds, one should expec ha
eaching ope a ion bandwid hs in he megahe z ange, indeed,
cu en s a e o he a in hin- ilm piezoelec ic MEMS, can
achie e ~30MHz o swi ching equencies (54–56). In e ms o s a-
bili y and epea abili y o ope a ion, hin- ilm piezoelec ic MEMS
can su i e mo e han 1011 cycles a ull 20 V o ac cycles o s an-
da d ope a ing condi ions (23°C, 35% ela i e humidi y), d i ing
by ~10% du ing i s li e ime (61), al hough he epea abili y wi h-
in ~1nm is easible wi h accu a e posi ion eedback by, e.g., op ical,
capaci i e, o piezo esis i e sensing. As a as he ib a ion ins abili y
is conce ned, i is impo an ha he MEMS de ice and glass pla e
esonances a e no exci ed, which is usually he case once esonance
equencies a e abo e 1kHz. The cu en MEMS de ice has a eso-
nance equency o ~4kHz and ha o he glass pla e is much highe .
Consequen ly, no ib a ion is expec ed unde no mal ci cums ances
and no ins abili y was obse ed.
Concluding he p esen a ion o he demons a ed MEMS-OMS
o pola iza ion-independen dynamic beam s ee ing, we would like
o no e ha , al hough he expe imen ally obse ed pe o mance
(Fig.4) is somewha in e io o ha expec ed om ou simula ions
(Fig.2), he expe imen al pe o mance can be imp o ed. The de e-
io a ion can be a ibu ed pa ly o ab ica ion impe ec ions and
o he smalles ai gap a ha was achie ed in p ac ice. I seems ha
he ai gap dec eases wi h applying he ac ua ion ol age only up
o ~3.75 V, esul ing he eby in inc easing +1s and dec easing 0 h
o de di ac ion e iciencies, whe eas o la ge ol ages, he MEMS
Fig. 4. Pola iza ion-independen dynamic beam s ee ing: Cha ac e iza ion. (A) Op ical images a he di ec objec (DI) and Fou ie image (FI) planes o he e lec ed
ligh om MEMS-OMS unde ac ua ion ol ages o Va1 = 0.00 V ( op) and Va2 = 3.75 V (middle) o TM/TE no mally inciden ligh wi h 800-nm wa eleng h. Re lec ed ligh
om uns uc u ed subs a e (bo om) in he MEMS-OMS de ice is also eco ded as a e e ence. (B) Di ac ion e iciencies o di e en o de s (|m| ≤ 1) measu ed as a unc ion
o he ac ua ion ol age o TM/TE inciden ligh wi h 800-nm wa eleng h. (C) Di ac ion e iciencies o di e en o de s (|m| ≤ 1) measu ed as a unc ion o he wa eleng h
o TM/TE inciden ligh . (D) Response ime o he di e en di ac ion o de s (m = 0/+1) measu ed by ac ua ing he MEMS mi o wi h a pe iodic ec angle signal.
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
6 o 11
mi o s a s o mo e sligh ly away om he OMS, p obably because
o he esidual con aminan s on he subs a e o bending a he pedes al
edges ha p e en he MEMS mi o om mo ing u he close o
he OMS su ace. Bo h be e ab ica ion accu acy and smalle ai
gaps a e easible and expec ed o be ealized in u he expe imen s.
Pola iza ion-independen dynamic 2D ocusing: Design
The MEMS-OMS design o ealizing dynamically con olled
pola iza ion-independen 2D beam ocusing in e lec ion equi es
he choice o diame e D and ocal leng h o he OMS lens ha , in
u n, de e mines he nume ical ape u e (NA) o he gi en e ac-
i e index in he image space n=1.46 a an inciden wa eleng h o
=800 nm: NA=nsin[ an−1(D/2 )]. To ealize s ong ocusing, we
chose D=14 m and =15m, so ha NA ≈ 0.62 is expec ed,
which should be adequa e o enable high-e iciency e lec i e 2D
ocusing (7). Following he same design app oach used in demon-
s a ing MEMS-OMS o dynamic beam s ee ing, we use he phase
esponse calcula ed wi h ai gap a=20nm o di e en nanob ick
leng hs ( he ed dashed line in Fig.2C) o ex ac he p ope uni
cells and a ange hem in o a ci cula egion wi h D=14m
(Fig.5A), app oxima ing a hype boloidal phase p o ile (7,9)
2D = 2
_
n( −
√
_
x 2 + y 2 + 2 ) in he xy plane (Fig.5B). The abo e
phase p o ile is also disc e ized wi h he s ep size =250nm along
bo h x and y di ec ions, ma ching he uni cell size (=250 nm). In
con as o he p e ious wo k, we do no limi he choice o uni
cells o a disc e e design space [i.e., uni cells wi h disc e e phase
s eps o 45° (7)]. Ins ead, app op ia e leng hs o he nanob icks a e
chosen om he en i e space o simula ion esul s ( he ed dashed
line in Fig.2C), hus ensu ing be e sampling o he 2D phase p o-
ile wi h mino de ia ions ( ig. S7, A and B) om he equi ed one
(Fig.5B). The de ia ion be ween he equi ed and a ailable phase
p o iles esul s mos ly om he achie able phase co e age o ~270°,
a limi a ion ha could be ci cum en ed by including mo e complex
uni cell elemen s such as de uned GSP esona o s (62) ha can also
be cons uc ed squa e-like o ensu e he pola iza ion-independen
ope a ion o by using c oss-like nanob icks, allowing o a wide
phase co e age (57).
Bea ing in mind high compu a ional demands when simula ing
2D ocusing (and hus ape iodic) OMSs, we es ima e he ocusing
pe o mance by simula ing he co esponding ( educed o a 1D
ape iodic con igu a ion) OMS ( ig. S7, C and D), which is designed
Fig. 5. Pola iza ion-independen dynamic 2D ocusing: Design. (A) Top iew o he OMS designed o dynamic 2D ocusing. (B) The phase p o ile equi ed o ocus
adia ion wi h ocal leng h o 15 m a 800-nm wa eleng h. (C and D) Dis ibu ions o he e lec ed in ensi y o TM inciden ligh wi h 800-nm wa eleng h a ai gaps o
a = 20 and 350 nm, espec i ely. (E and F) Dis ibu ions o he e lec ed TM elec ic ield (x componen ) a 800-nm wa eleng h o ai gaps o a = 20 and 350 nm, espec i ely.
(G and H) Focusing e iciencies calcula ed as a unc ion o he ope a ing wa eleng h and ai gap a o TM/TE pola iza ions. The g een, black, and cyan lines indica e he
cases o = 750, 800, and 950 nm, espec i ely. (I) Focusing e iciencies calcula ed as a unc ion o he ai gap a o TM/TE pola iza ions wi h espec i e 750-, 800-, and
950-nm wa eleng hs.
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
7 o 11
o p o ide a 1D hype boloidal phase p o ile 1D = 2
_
n( −
√
_
x 2 + 2 ) ,
while he D, , and a e he same as ha o he abo e-designed OMS
wi h he 2D phase p o ile. The e lec ed in ensi y dis ibu ions
calcula ed o his simpli ied MEMS-OMS unde TM/TE inciden
ligh a 800-nm wa eleng h wi h a=20nm mani es high ocusing
quali y wi h a di ac ion-limi ed spo si ua ed a he ocal leng h
o ~15 m (Fig.5C and ig. S7E). Fo inc eased ai gaps, he phase
g adien s p oduced by nanob icks wi h di e en leng hs p og essi ely
dec ease (Fig.2,BandC), app oaching ze o a an ai gap o 350 nm,
wi h he e lec ion ans o med in o specula e lec ion (Fig.5D and
ig. S7F). The associa ed e lec ed elec ic ields calcula ed nea he
ocus display smoo hly con e ging and plana wa e on s a ai gaps
a=20nm (Fig.5E and ig. S7G) and 350nm (Fig.5F and ig. S7H),
espec i ely, implying a high-e iciency ope a ion o he ac ua ed
MEMS-OMS. Taking in o accoun he possibili y o adjus ing he ai
gap o maximize he ocusing e iciency a o he ( han he design)
wa eleng hs, we e alua ed he ocusing e iciencies achie able a
di e en wa eleng hs wi h a ied ai gaps (Fig.5,GandH). The
maximum achie able ocusing e iciencies a he design wa eleng h
o 800nm a e es ima ed o be ~64/66% (TM/TE) o he ai gap
o ~20nm as expec ed. Fo o he wa eleng hs, he pola iza ion-
independen ocusing beha io is well main ained, while he co e-
sponding maximal ocusing e iciencies a e expec ed o achie e a
sligh ly di e en ai gaps. To be e isualize his ea u e, he ocusing
e iciency as a unc ion o he ai gap is explici ly plo ed o dis inc
wa eleng hs o 750, 800, and 950nm (Fig.5I), showing o all wa e-
leng hs a nea ly linea dec ease o he e iciency o inc easing ai
gaps wi hou no iceable changes in he e lec ed ield dis ibu ions
( ig. S7, I o L).
Pola iza ion-independen dynamic 2D
ocusing: Cha ac e iza ion
The MEMS-OMS o pola iza ion-independen dynamic e lec i e
2D ocusing designed as desc ibed abo e (Fig.5) was assembled ol-
lowing he ab ica ion and p echa ac e iza ion p ocesses simila o
hose used when assembling he dynamic beam s ee ing MEMS-OMS.
Op ical mic oscopy and SEM a e used o moni o ing he possible
con aminan s on he OMS su ace and he ab ica ion quali y ( he
uppe -le inse o Fig.6A and ig. S8).
To cha ac e ize he dynamic ocusing MEMS-OMS, we elec i-
cally ac ua ed he MEMS mi o and obse ed co esponding op ical
esponses in he di ec objec plane (Fig.6B). Since he MEMS-
OMS was designed o exhibi a e y sho ocal leng h o ~15 m, i
was no possible o di ec ly access he ocal plane using a beam spli e
(BS) and a low-di e gen inciden lase beam. Ins ead, he ocusing
e ec was e i ied by illumina ing he MEMS-OMS wi h a ocused
inciden beam and placing he MEMS-OMS su ace plane B a a
dis ance o ~2 ( he double ocal leng h o he MEMS-OMS) away
om he inciden beam ocal plane A (see inse in Fig.6A). Acco ding
o he ay op ics, he beam e lec ed by he OMS (when close o he
MEMS mi o ) will hen be ocused again a he ocal plane o he
objec i e (plane A in he bo om- igh inse o Fig.6A). A he same
ime, he e lec ion om he uns uc u ed subs a e su ace (ou side
he OMS a ea) would be s ong di e ging (see he bo om- igh inse
in Fig.6A). I one mo es he MEMS-OMS su ace o plane A, hen
he e lec ion beha io will be e e sed: The e lec ion by he OMS
will be di e ging (a e he objec i e) and he e lec ion by he un-
s uc u ed su ace collima ed. This p ocedu e was success ully
used and desc ibed in de ail in he p e ious expe imen conduc ed
wi h he s a ic ocusing OMS (7). In he cu en case wi h he
dynamic ocusing MEMS-OMS, i is expec ed o he MEMS-OMS
a angemen o swi ch be ween he ocusing con igu a ion, when
he applied ol age would b ing he OMS e y close o he MEMS
mi o , and he mi o e lec ing con igu a ion o ela i ely small
applied ol ages ha would co espond o su icien la ge OMS
and MEMS mi o sepa a ions.
To obse e his ans o ma ion, we moni o ed he e lec ed ligh
om he MEMS-OMS posi ioned a plane B while ac ua ing he
MEMS mi o . Fo bo h pola iza ions, he swi ching o he e lec ed
ligh be ween he mi o (a Vb1=10.00 V) and ocusing (a
Vb2=14.50 V) cases was clea ly isualized (Fig.6B), wi h he ocus-
ing e iciencies eaching hei maxima o ~56/53% a Vb2=14.50 V
o he espec i e TM/TE ligh incidence a he wa eleng h o
800nm (Fig.6A). A he same ime, he e lec ion om he uns uc-
u ed subs a e su ace was no in luenced wi h he applied ol ages,
e ealing, howe e , ha he e lec ion om he subs a e a plane A
is no ably simila o he TM/TE e lec ion om he OMS a plane B,
wi h he applied ol age being Vb2=14.50 V (Fig.6B). The la e
e idences a a he high e iciency and excellen quali y o pola iza ion-
independen ocusing by he MEMS-OMS a Vb2=14.50 V. The
dynamic e olu ion o he e lec ed ield om he MEMS-OMS
posi ioned a plane B, induced by ac ua ing he MEMS mi o wi h
s epwise inc eased ol ages om 10.00 o 14.50 V, is clea ly obse ed
wi h a CCD came a (mo ie S2). Because o he usage o he same
Fig. 6. Pola iza ion-independen dynamic 2D ocusing: Cha ac e iza ion.
(A) Focusing e iciencies measu ed as a unc ion o he ac ua ion ol age o TM/TE
inciden ligh wi h 800-nm wa eleng h. The uppe -le inse is a ypical SEM image
o he OMS ep esen ing 14-m-diame e and 250-nm-pe iod a ay o di e en ly
sized gold nanob icks designed o dynamic 2D ocusing. Scale ba , 2 m. The
bo om- igh inse illus a e he measu emen me hod in which he inciden beam
is ocused a plane A ( ocal plane o he objec i e) and impinging on he uns uc u ed
subs a e o OMS a ea o he MEMS-OMS a plane B (2 dis ance away om he ocal
plane o he objec i e), esul ing in espec i e di e gen o ocused e lec ed ields.
(B) Op ical images o he e lec ed ligh om he uns uc u ed subs a e and OMS
a ea o he MEMS-OMS posi ioned a plane B wi h ac ua ion ol ages o Vb1 = 10.00 V
and Vb2 = 14.50 V o TM/TE inciden ligh a 800-nm wa eleng h. The e lec ed
ligh om he uns uc u ed subs a e and OMS a ea o he MEMS-OMS posi ioned
a plane A was also eco ded as a e e ence.
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
8 o 11
MEMS componen as ha in he dynamic beam s ee ing MEMS-OMS,
simila esponse ime o ~0.4 ms is expec ed. I is las ly wo h
no ing ha , acco ding o he cu en s a e o he a in hin- ilm
piezoelec ic MEMS echniques (54–56), MEMS-OMS componen s
wi h a ew megahe z o swi ching bandwid h should be easible and
expec ed o u he de elopmen s.
DISCUSSION
We ha e de eloped he elec ically d i en dynamic MEMS-OMS
pla o m by combining a hin- ilm piezoelec ic MEMS mi o wi h
a GSP-based OMS. This pla o m o e s con ollable phase and
ampli ude modula ion o he e lec ed ligh by inely ac ua ing he
MEMS mi o . We ha e designed and expe imen ally demons a ed
MEMS-OMS de ices ope a ing in he nea -in a ed wa eleng h
ange o dynamic pola iza ion-independen beam s ee ing and
e lec i e 2D ocusing, bo h exhibi ing e icien (~50%), b oadband
(~20% nea he ope a ing wa eleng h o 800 nm), and as (<0.4 ms)
ope a ion. No e ha he ope a ion bandwid h can be ma kedly in-
c eased when using he ci cula ly pola ized ligh whose ans o ma ion
elies on he OMS, making use o he geome ical (Pancha a nam-
Be y) phase (17). The ope a ion o bo h de ices elies on he phase
esponse ans o ma ion when changing he MEMS-OMS sepa a-
ion by adjus ing he applied ol age wi hin he ange o ~4 V. The
same ope a ion p inciple can be used o design a MEMS-OMS o
dynamically con olling any unc ionali y a ailable o con en ional
GSP-based OMSs, om pola iza ion con ol/de ec ion o ec o /
o ex beam gene a ion (59): Fo a gi en smalles ai gap, one
designs he GSP-based OMS exhibi ing a equi ed unc ionali y ha
can hen be swi ched on and o by mo ing he MEMS mi o
owa d and away om he OMS su ace.
Mo eo e , he non i ial modi ica ion o he size-dependen phase
esponse wi h he MEMS-OMS sepa a ion (Fig.2B), which can
accu a ely be adjus ed by elec ical MEMS ac ua ion, sugges s a
possibili y o ealizing mo e sophis ica ed dynamic unc ionali ies.
One unc ionali y o pa icula in e es o comme cial applica ions
is he possibili y o swi ching be ween mul iple di ac ion o de s o
allow o quasi-con inuous beam s ee ing ( o use in, e.g., LIDAR
applica ions). Thus, we ha e also designed and expe imen ally
demons a ed he MEMS-OMS de ice o pola iza ion-independen
dynamic beam s ee ing be ween h ee (0 h, 1s , and 2nd) di ac ion
o de s, co esponding o e lec ion angles o 0°, 5.2°, and 10.5° in
glass (i.e., 0°, 7.7°, and 15.5° in ai ) unde no mally inciden ligh
wi h 800-nm wa eleng h. The OMS con igu a ion ( ig. S9) consis ed
o wo OMSs wi h di e en supe cells wi h sc1=12 and sc2=
24 op imized a wo dis inc ai gaps and in e lea ed by adop ing
a andom-in e lea ing s a egy (63). The expe imen al cha ac e i-
za ion ( ig. S10) has con i med he in ended dynamic MEMS-OMS
esponse: Wi h he ac ua ion ol age inc easing, he +1s and +2nd
di ac ion o de s became isible, succeeding one ano he , in acco d-
ance wi h ou simula ions ( ig. S9, K and L).
Ano he p omising di ec ion o u he esea ch and de elop-
men is o ci cum en he necessi y o b inging he MEMS mi o
e y close (~100 nm) o he OMS su ace. Fo la ge MEMS-OMS
sepa a ions, one can make use o localized plasmon esonances
because o exci a ion o sho - ange su ace plasmon pola i ons
(SR-SPPs) in hin me al ilms (58). Ou p elimina y simula ions
showed ha he SR-SPP esonances hyb idize wi h he Fab y-Pé o
esonances (suppo ed wi h wa eleng h-la ge ai gaps) (64,65) and
open a simila o he conside ed abo e ou e o modi y he OMS
phase esponse by con olling he ai gap ( ig. S11, A o C). No e ha
a ce ain ai gaps (sepa a ed by hal o he wa eleng h), he e lec ed
phase becomes independen on he nanob ick size ( ig. S11, A and
B), esul ing he eby in he mi o -like beha io ( ig. S11, D and E).
In be ween hese ai gaps, he e a e gaps a which he phase does
depend on he nanob ick size (see a dashed line a he gap o
1250nm in ig. S11B). A hese gaps, he nanob ick sizes can be
chosen in a manne enabling one o ealize a phase-g adien me a-
su ace ( ig. S11, F and G). Swi ching be ween hese wo dis inc ai
gaps esul s he e o e in swi ching be ween he mi o -like and g a-
dien me asu ace beha io , which is simila o swi ching be ween
he same ypes o esponses o he dynamic GSP-based me asu aces.
Wi h his app oach, he MEMS-OMS can be ope a ed nea he ai
gap o ~1 m o mo e, as demons a ed wi h ou simula ions o
dynamic beam s ee ing ( ig. S11, D o L), hus a oiding he p oblem
o ealizing nanome e -sized ai gaps. O e all, we belie e ha di e se
unc ionali ies wi h dynamically econ igu able pe o mances can be
ealized using he de eloped MEMS-OMS pla o m, hus opening
ascina ing pe spec i es o success ul ealiza ion o high-pe o mance
dynamically con olled de ices wi h po en ial applica ions in u u e
econ igu able/adap i e op ical sys ems.
MATERIALS AND METHODS
Simula ion me hods
All nume ical simula ions we e pe o med using COMSOL Mul i-
physics 5.5. We modeled one indi idual glass-Au-ai -Au uni cell
(Fig.2A), whe e pe iodic bounda y condi ions we e applied in bo h
x and y di ec ions, and linea ly x-pola ized ligh a he design wa e-
leng h o 800nm was no mally inciden on o he uni cell om he
uppe glass laye . The pe mi i i y o Au is desc ibed by he in e po-
la ed expe imen al alues (66), and he glass laye is aken as a lossless
dielec ic wi h a cons an e ac i e index o 1.46. Then, he complex
e lec ion coe icien s (Fig.2B) we e calcula ed as a unc ion o
nanob ick leng hs Lx, and ai gap a wi h o he pa ame e s being as
ollows: =800 nm, m=50 nm, =250 nm, and Ly=Lx o ensu e
he pola iza ion-independen op ical esponses.
To design he MEMS-OMS o dynamic beam s ee ing, he phase
esponse calcula ed wi h he ai gap a=20nm o di e en nano-
b ick leng hs is used o selec he leng hs o 12 nanob icks (Fig.2C)
o app oxima ing he e lec ion coe icien o an ideal blazed g a ing:
(x)=Aexp(i2x/sc) (6,8,57), whe e A≤1 is he e lec ion ampli-
ude, and sc=12 is he g a ing (supe cell) pe iod. Re lec ed ligh
di ec ed o di e en di ac ion o de s a e moni o ed, wi h di e en
ai gaps a and inciden wa eleng hs o es ima ing he dynamic
di ac ion e iciencies and ope a ion op ical bandwid hs, espec i ely
(Fig.2,D o I, and ig. S2, F o L). He e, he di ac ion e iciencies
a e de ined as he a ios o he ligh in ensi ies (in glass) in he co e-
sponding di ac ion o de s o he inciden (in glass) ligh in ensi y.
MEMS-OMS o dynamic beam ocusing is designed and simu-
la ed in a simila ashion. Nanob icks om he phase esponse cal-
cula ed wi h he ai gap a=20nm o di e en nanob ick leng hs
(Fig.2C) a e selec ed o app oxima e a 1D hype boloidal phase p o ile
o 1D = 2
_
n( −
√
_
x 2 + 2 ) (7,9) wi hin a 14-m-diame e egion
in he xy plane ( ig. S7, C and D). Re lec ed ields a e moni o ed o
isualize he dynamic beam ocusing and es ima e co esponding
ocusing e iciencies as a unc ion o he gap sizes a and inciden
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
9 o 11
wa eleng hs , o bo h TM and TE pola iza ions (Fig.5,C o I, and
ig. S7, E o L). He e, he ocusing e iciencies a e de ined as he a io
o he ligh powe om he co esponding ocal spo (in glass) o he
inciden (in glass) ligh powe . No e ha bo h di ac ion and ocus-
ing e iciencies ob ained in ou simula ions should no be di ec ly
compa ed o he expe imen al alues ha we e measu ed in ai
because o he e lec ions a he glass-ai in e ace. Conside ing he
ac ha all op ical ields p opaga e a di ec ions close o he no mal
o he OMS su ace and dis ega ding mul iple e lec ions, one can
es ima e he expec ed di e ence be ween he quan i ies ob ained
o he ields in ai and in glass as ai ≈0.93glass, i.e., he di e ence
amoun s o ~7%.
Fab ica ion and assembly o he MEMS-OMS de ices
The OMSs o de eloping MEMS-OMS o dynamic beam s ee ing/
ocusing we e ab ica ed using s anda d elec on beam li hog aphy
(EBL), hin- ilm deposi ion, and li -o echniques. Fi s , a 100-nm- hick
poly(me hyl me hac yla e) (2% in anisole; Mic oChem) laye and a
40-nm- hick conduc i e polyme laye (AR-PC 5090, All esis ) we e
successi ely spin-coa ed on a 16mm by 16mm glass subs a e
(Bo o loa 33 wa e , Wa e Uni e se). No e ha he glass subs a e
was p ep ocessed o ha e a 10-m-high ci cula /c oss-shaped
pedes al on one side using op ical li hog aphy and we e ching. The
OMSs we e hen de ined on he pedes al o he glass subs a e using
EBL (JEOL JSM-6500F ield-emission SEM wi h a Rai h Elphy
Quan um li hog aphy sys em) and subsequen ly de eloped in 1:3
solu ion o me hyl isobu yl ke one and isop opyl alcohol. A e
de elopmen , a 1-nm Ti adhesion laye and a 50-nm Au laye we e
deposi ed using he mal e apo a ion. The Au nanob icks we e las ly
o med a op he pedes al on he glass subs a e a e a li -o p o-
cess (Fig.3 and ig. S8). Owing o he la ge size o he MEMS mi o
(~3mm in diame e ) in compa ison o he OMS (30 m by 30 m
in size), he pedes al on he glass subs a e is e y p ac ical o
educing he possible con aminan s be ween he MEMS mi o and
OMS su ace, hus p omising high-e iciency modula ion o he
MEMS-OMS de ices.
The MEMS mi o , which is e y simila o he p e iously e-
po ed ul aplana , long-s oke, and low- ol age piezoelec ic mic o-
mi o (53), is ab ica ed using s anda d semiconduc o manu ac u ing
p ocesses ( ig. S4A) and inco po a ing hin- ilm lead zi cona e i ana e
(PZT) o ac ua ion. Fi s , a pla inum-bo om elec ode, a 2-m- hick
PZT ilm, and a op elec ode consis ing o TiW/Au we e deposi ed
on a SOI wa e ( ig. S4A, i s panel). Then, a cen al ci cula ape -
u e o 3mm was opened by using deep eac i e ion e ching o silicon
and e ching o he bu ied oxide ( ig. S4A, second panel). An annulus
ench is e ched in o he backside o he wa e , he eby eleasing he
ci cula pla e. Las , he wa e backside is spu e ed wi h Au ( ig. S4A,
hi d panel) o ac ing as he ul a la MEMS mi o ha is o i al
impo ance in de eloping dynamic MEMS-OMS.
A e he ab ica ion o bo h OMS and MEMS mi o , we mo e
o he assembly and packaging p ocesses o making MEMS-OMS
de ices ( ig. S4A, ou h panel). Be o e assembly, he su ace opog-
aphy o he MEMS mi o and glass subs a e we e measu ed by a
whi e ligh in e e ome y (Zygo NewView 6000), so as o selec
a o able a eas on bo h sides wi h he leas amoun o con aminan s
and su ace oughness ha migh obs uc he MEMS mi o om
ge ing close enough o he OMS. Then, he MEMS mi o was glued
o he glass subs a e upon which he OMS has been s uc u ed ( ig. S4,
B o D). Ge ing he mi o and OMS pa allel was done by adjus ing
he il o he MEMS mi o using he piezoelec ic elec odes. The
spacing be ween he MEMS mi o and he glass subs a e ( a+ m)
was measu ed o be commonly ~2m a e moun ing ( ig. S4E),
well wi hin he 6-m mo ing ange o he MEMS mi o s ( ig. S4F).
Las , he MEMS-OMS was glued o a p in ed ci cui y boa d, and
gold wi e bonding is used o connec elec ically o he MEMS elec-
odes o enabling simple connec ion o a ol age con olle used
o ac ua e he MEMS mi o .
A e he MEMS-OMS assembly, we applied mul iwa eleng h
in e e ome y o es ima e he smalles achie able sepa a ion be ween
he MEMS mi o and OMS su ace ( ig. S5). We ound by ac ua ing
he MEMS mi o ha , o se e al assemblies, his gap ( m+ a) can
be as small as ~100nm co esponding o a~50 nm, and hese samples
we e hen selec ed o u he op ical cha ac e iza ions.
Op ical cha ac e iza ion o MEMS-OMS
To cha ac e ize he pe o mances o he MEMS-OMS o dynamic
2D wa e on shaping, we used a ibe -coupled wa eleng h- unable
Ti:sapphi e lase (Spec a-Physics 3900S; wa eleng h ange, 700 o
1000 nm), whose ligh was di ec ed h ough a hal -wa e pla e
(AHWP05M-980, Tho labs), a Glan-Thompson pola ize and a
i s BS (BS1; BS014, Tho labs) successi ely, and hen ocused by an
objec i e (Obj, M Plan Apo, ×20/×50 magni ica ions; Mi u oyo) on o
he MEMS-OMSs. The e lec ed ligh was collec ed by he same ob-
jec i e and di ec ed ia BS1 and a second BS (BS2; BS014, Tho labs) o
wo op ical pa hs e mina ed wi h wo CCD came as (DCC1545M,
Tho labs) o isualizing espec i e di ec objec and Fou ie plane
images ( ig. S6). No e ha he objec i e o ×20/0.42 and ×50/0.55
a e used o measu ing espec i e MEMS-OMS o dynamic beam
s ee ing and ocusing.
Du ing he measu emen , he MEMS mi o was elec ically ac-
ua ed o modula e he op ical esponses o he MEMS-OMS de ices.
To cha ac e ize he MEMS-OMS o dynamic beam s ee ing, we
measu ed bo h di ac ion e iciencies and esponse ime wi h he
expe imen al se up shown in ig. S6. Fo es ima ing he di ac ion
e iciencies, we eco ded he in ensi y o spa ially sepa a ed 0 h/±1s
di ac ion o de s using a CCD came a a he Fou ie plane o he
lase beam being on he OMS a ea, which is hen no malized wi h he
e lec ion in ensi y om an uns uc u ed subs a e in he MEMS-OMS
componen s. The esponse ime o he MEMS-OMS was e alua ed
by ac ua ing he MEMS mi o wi h a pe iodic ec angle signal om
a unc ion gene a o (TOE 7402, TOELLNER). The spa ially sepa-
a ed di ac ion o de s a he Fou ie plane could be selec ed by an i is
and hen p ojec ed o a pho ode ec o (PDA20CS-EC, Tho labs),
which was connec ed o an oscilloscope (DSOX2024A, Keysigh )
o isualizing and eco ding he co esponding modula ed signals.
In he esponse ime measu emen , we eco ded he 0 h/+1s di ac-
ion o de s o he MEMS-OMS componen s, showing o e all good
epea abili y and s abili y o he ac ua ed MEMS-OMS componen s
wi h he pe iodically elec ical signals (Fig.4D).
SUPPLEMENTARY MATERIALS
Supplemen a y ma e ial o his a icle is a ailable a h p://ad ances.sciencemag.o g/cgi/
con en / ull/7/26/eabg5639/DC1
REFERENCES AND NOTES
1. N. Yu, F. Capasso, Fla op ics wi h designe me asu aces. Na . Ma e . 13, 139–150 (2014).
2. D. Lin, P. Fan, E. Hasman, M. L. B onge sma, Dielec ic g adien me asu ace op ical
elemen s. Science 345, 298–302 (2014).
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
10 o 11
3. H.-T. Chen, A. J. Taylo , N. Yu, A e iew o me asu aces: Physics and applica ions. Rep.
P og. Phys. 79, 076401 (2016).
4. H.-H. Hsiao, C. H. Chu, D. P. Tsai, Fundamen als and applica ions o me asu aces. Small
Me hods 1, 1600064 (2017).
5. F. Ding, A. Po s, S. I. Bozhe olnyi, G adien me asu aces: A e iew o undamen als
and applica ions. Rep. P og. Phys. 81, 026401 (2018).
6. S. Sun, K.-Y. Yang, C.-M. Wang, T.-K. Juan, W. T. Chen, C. Y. Liao, Q. He, S. Xiao, W.-T. Kung,
G.-Y. Guo, L. Zhou, D. P. Tsai, High-e iciency b oadband anomalous e lec ion by
g adien me a-su aces. Nano Le . 12, 6223–6229 (2012).
7. A. Po s, M. G. Nielsen, R. L. E iksen, S. I. Bozhe olnyi, B oadband ocusing la mi o s
based on plasmonic g adien me asu aces. Nano Le . 13, 829–834 (2013).
8. A. Po s, O. Alb ek sen, I. P. Radko, S. I. Bozhe olnyi, Gap plasmon-based me asu aces
o o al con ol o e lec ed ligh . Sci. Rep. 3, 2155 (2013).
9. M. Kho asaninejad, W. T. Chen, R. C. De lin, J. Oh, A. Y. Zhu, F. Capasso, Me alenses a
isible wa eleng hs: Di ac ion-limi ed ocusing and subwa eleng h esolu ion imaging.
Science 352, 1190–1194 (2016).
10. Y. Zhao, A. Alù, Manipula ing ligh pola iza ion wi h ul a hin plasmonic me asu aces.
Phys. Re . B. 84, 205428 (2011).
11. A. A babi, Y. Ho ie, M. Baghe i, A. Fa aon, Dielec ic me asu aces o comple e con ol
o phase and pola iza ion wi h subwa eleng h spa ial esolu ion and high ansmission.
Na . Nano echnol. 10, 937–943 (2015).
12. Y. Kan, S. K. H. Ande sen, F. Ding, S. Kuma , C. Zhao, S. I. Bozhe olnyi, Me asu ace-
enabled gene a ion o ci cula ly pola ized single pho ons. Ad . Ma e . 32, 1907832 (2020).
13. F. Ding, B. Chang, Q. Wei, L. Huang, X. Guan, S. I. Bozhe olnyi, Ve sa ile pola iza ion
gene a ion and manipula ion using dielec ic me asu aces. Lase Pho onics Re . 14,
2000116 (2020).
14. E. Ka imi, S. A. Schulz, I. De Leon, H. Qassim, J. Upham, R. W. Boyd, Gene a ing op ical
o bi al angula momen um a isible wa eleng hs using a plasmonic me asu ace. Ligh
Sci. Appl. 3, e167 (2014).
15. Y. Yang, W. Wang, P. Moi a, I. I. K a chenko, D. P. B iggs, J. Valen ine, Dielec ic
me a- e lec a ay o b oadband linea pola iza ion con e sion and op ical o ex
gene a ion. Nano Le . 14, 1394–1399 (2014).
16. R. C. De lin, A. Amb osio, N. A. Rubin, J. P. Bal hasa Muelle , F. Capasso, A bi a y spin- o–
o bi al angula momen um con e sion o ligh . Science 358, 896–901 (2017).
17. G. Zheng, H. Mühlenbe nd, M. Kenney, G. Li, T. Zen g a , S. Zhang, Me asu ace
holog ams eaching 80% e iciency. Na . Nano echnol. 10, 308–312 (2015).
18. X. Li, L. Chen, Y. Li, X. Zhang, M. Pu, Z. Zhao, X. Ma, Y. Wang, M. Hong, X. Luo, Mul icolo
3D me a-holog aphy by b oadband plasmonic modula ion. Sci. Ad . 2, e1601102 (2016).
19. W. T. Chen, K.-Y. Yang, C.-M. Wang, Y.-W. Huang, G. Sun, I.-D. Chiang, C. Y. Liao, W.-L. Hsu,
H. T. Lin, S. Sun, L. Zhou, A. Q. Liu, D. P. Tsai, High-e iciency b oadband me a-holog am
wi h pola iza ion-con olled dual images. Nano Le . 14, 225–230 (2014).
20. D. Wen, F. Yue, G. Li, G. Zheng, K. Chan, S. Chen, M. Chen, K. F. Li, P. W. H. Wong,
K. W. Cheah, E. Y. B. Pun, S. Zhang, X. Chen, Helici y mul iplexed b oadband me asu ace
holog ams. Na . Commun. 6, 8241 (2015).
21. C. V. Poul on, M. J. By d, P. Russo, E. Timu dogan, M. Khandake , D. Ve meulen,
M. R. Wa s, Long- ange LiDAR and ee-space da a communica ion wi h high-
pe o mance op ical phased a ays. IEEE J. Sel. Top. Quan um Elec on. 25, 1–8 (2019).
22. B. Schwa z, Mapping he wo ld in 3D. Na . Pho onics. 4, 429–430 (2010).
23. J. Hahn, H. Kim, Y. Lim, G. Pa k, B. Lee, Wide iewing angle dynamic holog aphic s e eog am
wi h a cu ed a ay o spa ial ligh modula o s. Op . Exp ess 16, 12372–12386 (2008).
24. A. Koma , R. Paniagua-Domínguez, A. Mi oshnichenko, Y. F. Yu, Y. S. Ki sha ,
A. I. Kuzne so , D. Neshe , Dynamic beam swi ching by liquid c ys al unable dielec ic
me asu aces. ACS Pho onics 5, 1742–1748 (2018).
25. S.-Q. Li, X. Xu, R. M. Vee il, V. Valuckas, R. Paniagua-Domínguez, A. I. Kuzne so ,
Phase-only ansmissi e spa ial ligh modula o based on unable dielec ic me asu ace.
Science 364, 1087–1090 (2019).
26. A. Lininge , A. Y. Zhu, J.-S. Pa k, G. Pale mo, S. Cha e jee, J. Boyd, F. Capasso, G. S angi,
Op ical p ope ies o me asu aces in il a ed wi h liquid c ys als. P oc. Na l. Acad. Sci.
U.S.A. 117, 20390–20396 (2020).
27. Q. Wang, E. T. F. Roge s, B. Gholipou , C.-M. Wang, G. Yuan, J. Teng, N. I. Zhelude ,
Op ically econ igu able me asu aces and pho onic de ices based on phase change
ma e ials. Na . Pho onics. 10, 60–65 (2016).
28. C. H. Chu, M. L. Tseng, J. Chen, P. C. Wu, Y.-H. Chen, H.-C. Wang, T.-Y. Chen, W. T. Hsieh,
H. J. Wu, G. Sun, D. P. Tsai, Ac i e dielec ic me asu ace based on phase-change medium.
Lase Pho onics Re . 10, 986–994 (2016).
29. M. Zhang, M. Pu, F. Zhang, Y. Guo, Q. He, X. Ma, Y. Huang, X. Li, H. Yu, X. Luo, Plasmonic
me asu aces o swi chable pho onic spin–o bi in e ac ions based on phase change
ma e ials. Ad . Sci. 5, 1800835 (2018).
30. J. Tian, H. Luo, Y. Yang, F. Ding, Y. Qu, D. Zhao, M. Qiu, S. I. Bozhe olnyi, Ac i e con ol
o anapole s a es by s uc u ing he phase-change alloy Ge2Sb2Te5. Na . Commun. 10,
396 (2019).
31. X. Duan, S. T. Whi e, Y. Cui, F. Neub ech, Y. Gao, R. F. Haglund, N. Liu, Recon igu able
mul is a e op ical sys ems enabled by VO2 phase ansi ions. ACS Pho onics. 7,
2958–2965 (2020).
32. Y. Yao, R. Shanka , M. A. Ka s, Y. Song, J. Kong, M. Lonca , F. Capasso, Elec ically unable
me asu ace pe ec abso be s o ul a hin mid-in a ed op ical modula o s. Nano Le .
14, 6526–6532 (2014).
33. Z. Miao, Q. Wu, X. Li, Q. He, K. Ding, Z. An, Y. Zhang, L. Zhou, Widely unable e ahe z
phase modula ion wi h ga e-con olled g aphene me asu aces. Phys. Re . X 5, 041027
(2015).
34. B. Zeng, Z. Huang, A. Singh, Y. Yao, A. K. Azad, A. D. Mohi e, A. J. Taylo , D. R. Smi h,
H.-T. Chen, Hyb id g aphene me asu aces o high-speed mid-in a ed ligh modula ion
and single-pixel imaging. Ligh Sci. Appl. 7, 51 (2018).
35. M. C. She o , P. W. C. Hon, K. T. Foun aine, J. C. Ga cia, S. M. Pon i, V. W. B a , L. A. Swea lock,
H. A. A wa e , Expe imen al demons a ion o >230° phase modula ion in ga e- unable
g aphene-gold econ igu able mid-in a ed me asu aces. Nano Le . 17, 3027–3034
(2017).
36. J. an de G oep, J.-H. Song, U. Celano, Q. Li, P. G. Kik, M. L. B onge sma, Exci on esonance
uning o an a omically hin lens. Na . Pho onics. 14, 426–430 (2020).
37. Z. Li, K. Yao, F. Xia, S. Shen, J. Tian, Y. Liu, G aphene plasmonic me asu aces o s ee
in a ed ligh . Sci. Rep. 5, 12423 (2015).
38. P. C. Wu, R. A. Pala, G. Ka aie Shi manesh, W. H. Cheng, R. Sokhoyan, M. G ajowe ,
M. Z. Alam, D. Lee, H. A. A wa e , Dynamic beam s ee ing wi h all-dielec ic
elec o-op ic III–V mul iple-quan um-well me asu aces. Na . Commun. 10, 3654
(2019).
39. F. Neub ech, X. Duan, N. Liu, Dynamic plasmonic colo gene a ion enabled by unc ional
ma e ials. Sci. Ad . 6, eabc2709 (2020).
40. G. K. Shi manesh, R. Sokhoyan, P. C. Wu, P. C. Wu, H. A. A wa e , H. A. A wa e ,
Elec o-op ically unable mul i unc ional me asu aces. ACS Nano 14, 6912–6920
(2020).
41. J. Pa k, B. G. Jeong, S. I. Kim, D. Lee, J. Kim, C. Shin, C. B. Lee, T. O suka, J. Kyoung, S. Kim,
K.-Y. Yang, Y.-Y. Pa k, J. Lee, I. Hwang, J. Jang, S. H. Song, M. L. B onge sma, K. Ha,
S.-W. Hwang, H. Choo, B. L. Choi, All-solid-s a e spa ial ligh modula o wi h independen
phase and ampli ude con ol o h ee-dimensional LiDAR applica ions. Na . Nano echnol.
16, 69–76 (2021).
42. B.-W. Yoo, M. Megens, T. Sun, W. Yang, C. J. Chang-Hasnain, D. A. Ho sley, M. C. Wu, A 32
× 32 op ical phased a ay using polysilicon sub-wa eleng h high-con as -g a ing
mi o s. Op . Exp ess 22, 19029–19039 (2014).
43. Z. Han, K. Kohno, H. Fuji a, K. Hi akawa, H. Toshiyoshi, MEMS econ igu able
me ama e ial o e ahe z swi chable il e and modula o . Op . Exp ess 22, 21326–21339
(2014).
44. Z. Han, S. Colbu n, A. Majumda , K. F. Böh inge , MEMS-ac ua ed me asu ace Al a ez
lens. Mic osys . Nanoeng. 6, 79 (2020).
45. H.-S. Ee, R. Aga wal, Tunable me asu ace and la op ical zoom lens on a s e chable
subs a e. Nano Le . 16, 2818–2823 (2016).
46. M. L. Tseng, J. Yang, M. Semmlinge , C. Zhang, P. No dlande , N. J. Halas, Two-dimensional
ac i e uning o an aluminum plasmonic a ay o ull-spec um esponse. Nano Le . 17,
6034–6039 (2017).
47. L. Cong, P. Pi chappa, Y. Wu, L. Ke, C. Lee, N. Singh, H. Yang, R. Singh, Ac i e mul i unc ional
mic oelec omechanical sys em me ade ices: Applica ions in pola iza ion con ol,
wa e on de lec ion, and holog ams. Ad . Op . Ma e . 5, 1600716 (2017).
48. X. Zhao, J. Schalch, J. Zhang, H. R. Se en, G. Duan, R. D. A e i , X. Zhang, Elec omechanically
unable me asu ace ansmission wa epla e a e ahe z equencies. Op ica 5, 303–310
(2017).
49. A. She, S. Zhang, S. Shian, D. R. Cla ke, F. Capasso, Adap i e me alenses wi h
simul aneous elec ical con ol o ocal leng h, as igma ism, and shi . Sci. Ad . 4,
eaap9957 (2018).
50. E. A babi, A. A babi, S. M. Kamali, Y. Ho ie, M. S. Fa aji-Dana, A. Fa aon, MEMS- unable
dielec ic me asu ace lens. Na . Commun. 9, 812 (2018).
51. A. L. Hols een, A. F. Cihan, M. L. B onge sma, Tempo al colo mixing and dynamic beam
shaping wi h silicon me asu aces. Science 365, 257–260 (2019).
52. Z. Ren, Y. Chang, Y. Ma, K. Shih, B. Dong, C. Lee, Le e aging o MEMS echnologies
o op ical me ama e ials applica ions. Ad . Op . Ma e . 8, 1900653 (2020).
53. T. Bakke, A. Vogl, O. Że o, F. Tyhold , I.-R. Johansen, D. Wang, A no el ul a-plana ,
long-s oke and low- ol age piezoelec ic mic omi o . J. Mic omech. Mic oeng. 20,
064010 (2010).
54. Y. Qiu, J. V. Giglio i, M. Wallace, F. G iggio, C. E. M. Demo e, S. Coch an, S. T olie -McKins y,
Piezoelec ic mic omachined ul asound ansduce (PMUT) a ays o in eg a ed
sensing, ac ua ion and imaging. Senso s 15, 8020–8041 (2015).
55. Y. Lu, A. Heida i, D. A. Ho sley, A high ill- ac o annula a ay o high equency
piezoelec ic mic omachined ul asonic ansduce s. J. Mic oelec omech. Sys . 24,
904–913 (2015).
Downloaded om h ps://www.science.o g on Ap il 26, 2022
Meng e al., Sci. Ad . 2021; 7 : eabg5639 23 June 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
11 o 11
56. Y. Lu, D. A. Ho sley, Modeling, ab ica ion, and cha ac e iza ion o piezoelec ic
mic omachined ul asonic ansduce a ays based on ca i y SOI wa e s.
J. Mic oelec omech. Sys . 24, 1142–1149 (2015).
57. A. Po s, S. I. Bozhe olnyi, Plasmonic me asu aces o e icien phase con ol in e lec ion.
Op . Exp ess 21, 27438–27451 (2013).
58. J. Jung, T. Sønde gaa d, S. I. Bozhe olnyi, Gap plasmon-pola i on nano esona o s:
Sca e ing enhancemen and launching o su ace plasmon pola i ons. Phys. Re . B. 79,
035401 (2009).
59. F. Ding, Y. Yang, R. A. Deshpande, S. I. Bozhe olnyi, A e iew o gap-su ace plasmon
me asu aces: Fundamen als and applica ions. Nanopho onics 7, 1129–1156 (2018).
60. A. C. Lesina, D. Goodwill, E. Be nie , L. Ramunno, P. Be ini, On he pe o mance o op ical
phased a ay echnology o beam s ee ing: E ec o pixel limi a ions. Op . Exp ess 28,
31637–31657 (2020).
61. R. P. Dahl-Hansen, F. Tyhold , J. Gjessing, A. Vogl, P. Wi endo p, J. Vedum, T. Tybell, On
he e ec o wa e -induced deg ada ion o hin- ilm piezoelec ic
mic oelec omechanical sys ems. J. Mic oelec omech. Sys . 30, 105–115 (2021).
62. C. Damgaa d-Ca s ensen, F. Ding, C. Meng, S. I. Bozhe olnyi, Demons a ion o > 2 e lec ion
phase ange in op ical me asu aces based on de uned gap-su ace plasmon esona o s.
Sci. Rep. 10, 19031 (2020).
63. E. Maguid, I. Yule ich, D. Veksle , V. Kleine , M. L. B onge sma, E. Hasman, Pho onic
spin-con olled mul i unc ional sha ed-ape u e an enna a ay. Science 352, 1202–1206
(2016).
64. A. Be khou , A. F. Koende ink, Pe ec abso p ion and phase singula i ies in plasmon
an enna a ay e alons. ACS Pho onics 6, 2917–2925 (2019).
65. D. Ben Haim, L. Michaeli, O. A ayu, T. Ellenbogen, Tuning he phase and ampli ude
esponse o plasmonic me asu ace e alons. Op . Exp ess 28, 17923–17933 (2020).
66. P. B. Johnson, R. W. Ch is y, Op ical cons an s o he noble me als. Phys. Re . B 6,
4370–4379 (1972).
Acknowledgmen s
Funding: This p ojec has ecei ed unding om he ATTRACT p ojec unded by he EC unde
g an ag eemen 777222, he Uni e si y o Sou he n Denma k (SDU2020 unding), he VKR
Founda ion (Awa d in Technical and Na u al Sciences 2019 and g an nos. 00022988 and
37372), and he EU Ho izon 2020 esea ch and inno a ion p og am ( he Ma ie Skłodowska-Cu ie
g an ag eemen no. 713694). Au ho con ibu ions: C.D. and S.I.B. concei ed he idea. C.M.
and F.D. pe o med he nume ical simula ions. C.M. and C.W. ab ica ed he OMS samples,
P.C.V.T. and J.G. ab ica ed he MEMS mi o and assembled MEMS-OMS de ices, and C.M. and
P.C.V.T. conduc ed he op ical measu emen s. C.M. and M.T. pe o med he esponse ime
measu emen . All au ho s con ibu ed o he manusc ip w i ing. C.D. and S.I.B. supe ised he
p ojec . Compe ing in e es s: The au ho s decla e ha hey ha e no compe ing in e es s.
Da a and ma e ials a ailabili y: All da a needed o e alua e he conclusions in he pape a e
p esen in he pape and/o he Supplemen a y Ma e ials. Addi ional da a ela ed o his pape
may be eques ed om he au ho s.
Submi ed 15 Janua y 2021
Accep ed 11 May 2021
Published 23 June 2021
10.1126/sciad .abg5639
Ci a ion: C. Meng, P. C. V. Th ane, F. Ding, J. Gjessing, M. Thomaschewski, C. Wu, C. Di dal,
S. I. Bozhe olnyi, Dynamic piezoelec ic MEMS-based op ical me asu aces. Sci. Ad . 7, eabg5639
(2021).
Downloaded om h ps://www.science.o g on Ap il 26, 2022
and Video S5), as well as o di e en inciden LP s a es (θ
LP
=60°,
75°, 105°, 120°) and fixed DWP o ien a ion θ
DWP
=45° (Fig. 4c,
Supplemen a yFig.S14,andVideoS6).Thepola iza ion ajec o ies
on hePoinca ésphe ea eseen o
be di e en ly il ed o di e en DWP o ien a ions ha ing he
common poin , he inciden LP s a e |y>(Fig. 4b), while being
pa allel o he plane (S
1
,S
3
)and eflec ing he inciden LP s a e
(Fig. 4c). No e ha , o any gi en poin on he Poinca é sphe e, one
can iden i y sui able o ien a ions o he DWP and inciden LP s a e
enabling closed pola iza ion ajec o ies o pass his poin , so ha a
mul i ude o pola iza ion modula ion capabili ies can be ealized
wi h he same DWP. Finally, we would like o emphasize ha all
expe imen al esul s ag ee exceedingly well wi h he simula ions
wi hou any fi ing pa ame e s, demons a ing con incingly ha he
ab ica ed DWP beha es acco ding o he design and ha he
modeling app oach de eloped is well sui ed o use in he u u e
MEMS-OMS componen de elopmen s.
Discussion
To summa ize, capi alizing on ou de elopmen o he MEMS-OMS
pla o m22 we ha e demons a ed he elec ically d i en DWP
ope a ing in eflec ion wi h high pola iza ion con e sion e ficiencies
(~75%), b oadband ope a ion (~100 nm nea he ope a ing wa e-
leng h o 800 nm), as esponses (<0.4 ms) and ull- ange bi e in-
gence con ol ha enables comple ely enci cling he Poinca é sphe e
along ajec o ies de e mined by he inciden ligh pola iza ion and
DWP o ien a ion. I should be no ed ha , gi en he access o nm-
sized ai gaps, one can exploi he same design p inciple o ealize he
DWP wi h gap su ace plasmons being gene a ed22,a egime ha
p omises a b oade ope a ion wa eleng h ange (~160 nm) al hough
p obably a he expense o a lowe e ficiency (~50%)22.Impo an ly,
he gene al app oach de eloped can also be applied o design a DWP
ope a ing in ansmission by using a pa ially ansmi ing MEMS
mi o and placing an OMS in he middle o an FP ca i y25.Gi en
ha a mul i ude o pola iza ion modula ion capabili ies can be ea-
lized wi h he same DWP, we belie e ha he demons a ed elec-
ically d i en DWP configu a ion wi h ull- ange bi e ingence
con ol opens ascina ing pe spec i es o success ul in eg a ion o
high-pe o mance compac dynamic pola iza ion componen s in o
u u e minia u ized econfigu able/adap i e op ical ne wo ks and
sys ems26,27.
Me hods
Nume ical calcula ions. All nume ical simula ions we e done using COMSOL
Mul iphysics e sion 5.6. The model is composed o a ec angula olume wi h a
squa e oo p in wi h sides Λ=250 nm and pe iodic bounda y condi ions we e
employed o bo h uand di ec ions. The DWP uni cell is di ided in o wo pa s o
ai and glass, wi h one gold nanob ick placed agains he glass egion. The co ne s o
he nanob ick a e ounded wi h a 5 nm adius. The e ac i e index o ai is se as 1
and ha o glass as 1.46 o all wa eleng hs, while he gold pe mi i i y was in e -
pola ed as a unc ion o wa eleng h om expe imen al abula ed alues28.
Using his model, he complex eflec ion and ansmission coe ficien s o he
glass/OMS/ai in e ace a e calcula ed o bo h p opaga ion di ec ions (i.e., a
no mal inciden om glass o ai ) and o ligh linea ly pola ized along bo h uand
sepa a ely. These a e used o calcula e he o al eflec ion coe ficien FP by
including he gold subs a e wi h he FP equa ion19–21
FP ¼ 12 þ 12 21 23ei2kn2Ta
1 21 23ei2kn2Ta
ð1Þ
|y>
|x>
Vm
-8 -6 -4
|l>
| >
Vm
Time (ms)
-2 02468
0
6
12
In ensi y (mV)
0
6
12
14
17
20
Vol age Vm (V)
-4 -3 -2
Time (ms)
-1 01234
0
10
20
In ensi y (mV)
0
10
20
17
19
21
Vol age Vm (V)
ZWP HWP
c
dQWP TQWP QWP TQWP QWP TQWP QWP TQWP
ZWP HWP ZWP HWP ZWP HWP
b
| >
|l>
DI FI DI FI
DI FI DI FI
a
|x>
|y>
Vm(HWP)=17.1 V
Vm(TQWP)=14.4 VVm(QWP)=19.1 V
Vm(ZWP)=20.2 V
Fig. 3 Pola iza ion con e sion dynamics. a,bOp ical images o he eflec ed ligh a he di ec image (DI) and Fou ie image (FI) planes o he fixed inpu
linea pola iza ion (LP) s a e |y>(θ
LP
=90°) and dynamic wa e pla e (DWP) o ien a ion (θ
DWP
=45°) a λ=800 nm, showing pola iza ion modula ion
be ween o hogonal (a) LP s a es (|x>, |y>), co esponding o ze o-wa e pla e (ZWP)/hal -wa e pla e (HWP) ans o ma ion, by changing he ac ua ion
ol age V
m
om 20.2 o 17.1 V, and (b) ci cula pola iza ion (CP) s a es (| >, |l>), co esponding o qua e -wa e pla e (QWP)/ h ee-qua e s-wa e pla e
(TQWP) ans o ma ion, by changing he ac ua ion ol age V
m
om 19.1 o 14.4 V. Fo he FIs, he ligh is fil e ed using an i is le ing h ough only ligh
eflec ed om inside he o ange dashed ci cles (shown in DIs). The le mos column in (a,b) indica es he pola ize o ien a ions used o fil e ing he
espec i e pola iza ion s a es in he eflec ed ligh . Scale ba s in he DIs and FIs a e 10 μm and 0.02k
0
, espec i ely. c,dTempo al e olu ion o he eflec ed
ligh powe o cZWP/HWP and dQWP/TQWP ans o ma ions, measu ed by ac ua ing he MEMS mi o wi h a pe iodic ec angula signal and fil e ing
espec i e pola iza ion s a es.
NATURE COMMUNICATIONS | h ps://doi.o g/10.1038/s41467-022-29798-0 ARTICLE
NATURE COMMUNICATIONS | (2022) 13:2071 | h ps://doi.o g/10.1038/s41467-022-29798-0 | www.na u e.com/na u ecommunica ions 5
He e,
mn
(
mn
) deno es he eflec ion ( ansmission) coe ficien s o ligh inciden
on ma e ial n om ma e ial mand he ma e ials a e numbe ed 1, 2, 3 o
espec i ely glass subs a e, ai , and gold subs a e, n
2
ep esen s he e ac i e
index o he ai (i.e., medium 2 be ween he OMS laye and gold subs a e), and kis
he wa enumbe in a acuum. No e ha he e ec o he OMS on he in e ace
be ween he ai and glass subs a e is included in
12
,
21
,
12
, and
21
, and ha e en
o no mal incidence
FP
a e pola iza ion-dependen due o he aniso opy o he
OMS laye . The eflec ion coe ficien
23
om he ai /gold in e ace is calcula ed
di ec ly using he F esnel equa ion. An illus a ion o he DWP geome y is shown
in Supplemen a y Fig. S2 oge he wi h plo s explaining he beha io o he o al
eflec ion coe ficien
FP
wi h a ying T
a
.
Fo ai gaps much smalle han he wa eleng hs, he e is nea -field coupling
be ween he nanob icks and he gold subs a e in addi ion o he FP esonances1,20,
equi ing nume ical simula ions including also he gold subs a e in COMSOL. The
esul s ob ained using he FP equa ion we e confi med o gi e he same esul s as
COMSOL simula ions wi h he whole glass/nanob ick/ai /gold subs a e model
when T
a
> 100 nm o a wa eleng h o λ=800 nm.
As a final commen , o compa e wi h he measu emen s, he imagina y pa o
he gold pe mi i i y is inc eased by h ee imes in he simula ions o Figs. 2,4and
Supplemen a y Figs. S7, S9–S14, accoun ing o he su ace oughness and g ain
bounda y e ec s o he ab ica ed gold nanob icks as well as he inc eased damping
associa ed wi h he i anium (Ti) adhesion laye s be ween he gold/glass in e ace.
Fab ica ion. The OMS o de eloping MEMS-OMS DWP we e ab ica ed using
elec on-beam li hog aphy (EBL), hin-film deposi ion, and li -o echniques.
Fi s , a 100-nm- hick poly(me hyl me hac yla e) (PMMA A2, Mic oChem) laye
and a 40-nm hick conduc i e polyme laye (AR-PC 5090, All esis ) we e suc-
cessi ely spin-coa ed on he glass subs a e (Bo ofloa 33 wa e , Wa e Uni e se).
No e ha he glass subs a e was p ep ocessed o ha e a 10-μm-high ci cula
pedes al using op ical li hog aphy and we e ching, and he OMS pa e n was
defined on he pedes al using EBL (JEOL JSM-6500F field-emission SEM wi h a
Rai h Elphy Quan um li hog aphy sys em). A e de elopmen , he OMS laye was
o mula ed by deposi ing a 1-nm Ti adhesion laye and a 50-nm gold laye
(To nado 400, C yo ox) ollowed by li -o in ace one (Supplemen a y Fig. S4).
The pedes al on he glass subs a e is e y e ec i e o educing he possible
con aminan s be ween he MEMS mi o and OMS su ace, hus imp o ing he
s abili y and epea abili y o he DWP componen s. The MEMS mi o is ab i-
ca ed using s anda d semiconduc o manu ac u ing p ocesses (Supplemen a y
Fig. S3), in which hin-film lead zi cona e i ana e (PZT) is inco po a ed o long-
s oke, low- ol age elec ical ac ua ion. Fo use in he MEMS-OMS componen , he
ul a-fla MEMS mi o was spu e ed wi h a 100 nm gold laye . A e he gold
deposi ion, he MEMS mi o su ace is inspec ed wi h whi e ligh in e e ome y
(Zygo NewView 6300), showing o e all good fla ness and oughness all o e he
whole MEMS mi o (i.e., ~3 mm diame e ) (Supplemen a y Fig. S3).
The MEM-OMS-based DWP componen (Supplemen a y Fig. S4) was
assembled by gluing he MEMS mi o wi h he glass subs a e upon which OMS is
s uc u ed and hen glued o a p in ed ci cui boa d (PCB), ollowed by a gold wi e
bonding p ocess be ween he MEMS mi o and PCB o enabling simple elec ical
connec ion o a ol age con olle used o ac ua e he MEMS mi o .
Cha ac e iza ion. The expe imen al se up is shown in Supplemen a y Fig. S8. A
collima ed fibe -coupled supe con inuum lase (Supe K Ex eme, NKT) was
di ec ed h ough an HWP (AHWP10M-980, Tho labs), a mi o , a linea pola ize
(Pol
1
; LPNIR050-MP2, Tho labs), wo beam spli e s (BS
1,2
; CCM1-BS014, Tho -
labs) successi ely, and hen ocused on o he DWP samples by an objec i e (Obj; M
Plan Apo, ×20/0.42NA, Mi u oyo). The combina ion o HWP and Pol
1
is used o
al e ing he inpu LP s a es as well as he in ensi y. The eflec ed ligh was collec ed
by he same objec i e and passed h ough wo beam spli e s (BS
2,3
; CCM1-BS014,
Tho labs) and a ube lens (TL; TTL200-S8, Tho labs), gene a ing he fi s di ec
image plane whe e an i is is placed o fil e ing ou he eflec ed ligh wi hin he
DWP a ea. The fi s di ec image is hen ans o med by a elay lens (RL; AC254-
200-B-ML, =200 mm, Tho labs) o he co esponding Fou ie image and cap-
u ed by a CCD came a (CCD; DCC1545M, Tho labs), acco ding o a 2 config-
u a ion. No e ha a flip lens (FL; AC254-100-B-ML, =100 mm, Tho labs) is used
o swi ching be ween he di ec and Fou ie images, and a S okes analyze com-
posed o a QWP (AQWP10M-980, Tho labs) and a linea pola ize (Pol
2
;
LPNIR050-MP2, Tho labs) is implemen ed be o e he CCD came a o pe o ming
ull S okes pola ime y23. Two beam spli e s a e configu ed o c oss-
compensa ing he pola iza ion-dependen phase shi s in he beam spli e s o
bo h incidence and eflec ion ou es.
To ob ain he wa eleng h- esol ed ull S okes pa ame e s, we eplaced he CCD
came a wi h a fibe -coupled spec ome e (QE P o, Ocean Op ics) and conduc ed
measu emen s a he Fou ie image plane. By o a ing he QWP and Pol
2
,we
eco ded pola iza ion- esol ed spec a o I
x
(λ), I
y
(λ), I
a
(λ), I
b
(λ), I
(λ), I
l
(λ), and he
s okes pa ame e s (s
1
,s
2
,s
3
) a e calcula ed as s
1
=(I
x
(λ)–I
y
(λ))/(I
x
(λ)+I
y
(λ)),
s
2
=(I
a
(λ)–I
b
(λ))/(I
x
(λ)+I
y
(λ)), s
3
=(I
(λ)–I
l
(λ))/(I
x
(λ)+I
y
(λ)). Fo a
easonable compa ison wi h he simula ions, he s okes pa ame e s (s
1
,s
2
,s
3
) a e
no malized o he pola ized p opo ion o he eflec ed ligh beam: S1;2;3¼s1;2;3
DOP,
and he deg ee o pola iza ion (DOP) is defined as DOP ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
s2
1þs2
2þs2
3
p.The
pola iza ion con e sion e ficiency (Fig. 2and Supplemen a y Figs. S9–S11) is
defined as he a io o he eflec ed ligh powe in a specific pola iza ion channel
(i.e., |x>, |y>, |a>, |b>, | >, |l> ) o he inciden linea ly |y> pola ized ligh powe .
The coo dina es sys em used is indica ed in he lowe -le inse o Fig. 1a, wi h z
being he op ical axis, xand ya e ans e se axes in he labo a o y ame o
abc
θDWP=75°
θDWP=60°
θDWP=30°
θDWP=15°
θLP=60°
θLP=75°
θLP=105°
θLP=120°
y
u
45°
x
60° 75° 105° 120°
Inpu Pola iza ion:
DWP O ien a ion: θDWP=45°
Inpu Pola iza ion: |y> (θLP=90°)
DWP O ien a ion: θDWP=45°
Inpu Pola iza ion:
DWP O ien a ion:
|y> (θLP=90°)
y
x
30°
60° 75°
15°
u
u
u
u
Inc easing T
a
Inc easing Ta
Inc easing T
a
S1S2
S3
x( as )
y(slow)
QWP
z
x
u
45°
y
S1S2
S3
S1S2
S3
Fig. 4 Ve sa ile pola iza ion ans o ma ions. a–cCalcula ed (lines) and measu ed (ci cles) pola iza ion ajec o ies on he Poinca é sphe e ealized a he
wa eleng h o 800 nm by uning he ai gap T
a
o di e en dynamic wa e pla e (DWP) o ien a ions, illus a ing he di e si y o possible pola iza ion
ans o ma ions: acon inuous linea pola iza ion (LP) o a ion ealized by combining he DWP wi h a con en ional qua e -wa e pla e (QWP) wi h he
inciden LP s a e |y> and DWP o ien a ion θ
DWP
=45°, b a ious ellip ical pola iza ion ans o ma ions ealized o di e en DWP o ien a ions wi h he
fixed inciden LP s a e |y>, and c a ious ellip ical pola iza ion ans o ma ions ealized o di e en inciden LP s a es wi h he fixed DWP o ien a ion
θ
DWP
=45°. The g een s a s in (a–c) indica e espec i e inciden LP s a es.
ARTICLE NATURE COMMUNICATIONS | h ps://doi.o g/10.1038/s41467-022-29798-0
6NATURE COMMUNICATIONS | (2022) 13:2071 | h ps://doi.o g/10.1038/s41467-022-29798 -0 | www.na u e.com/na u ecommunica ions
e e ence, while uand a e ans e se axes o ien ed along he long and sho sides
o he ec angula nanob icks. The angle be ween uand xis deno ed θ
DWP
, while
he angle be ween he x-axis and he pola iza ion di ec ion o LP inciden ligh is
deno ed θ
LP
.
To es ima e he swi ching speeds be ween di e en o hogonal LP and CP bases (i.e.,
di e en DWP s a us), he se up desc ibed abo e is modified by eplacing he inpu
lase and CCD came a wi h a cw Ti:sapphi e lase (Spec a-Physics 3900 S, wa eleng h
ange: 700 o 1000 nm), and a pho ode ec o (PD; PDA20CS-EC, Tho labs),
espec i ely. The signals om he PD a e acqui ed wi h an oscilloscope (DSOX2024A,
Keysigh ). In he measu emen , he MEMS-OMS-based DWP is ac ua ed wi h
pe iodically al e na ing ol ages and di e en pola iza ions (i.e., |x>, |y>, | >and|l>)
can be fil e ed by he S okes analyze .
Da a a ailabili y
All da a ha suppo he findings o he s udy a e p o ided in he main ex and
Supplemen a y In o ma ion files. Raw da a a e a ailable om he co esponding au ho s
upon easonable eques .
Recei ed: 16 Decembe 2021; Accep ed: 31 Ma ch 2022;
Re e ences
1. Ding, F., Po s, A. & Bozhe olnyi, S. I. G adien me asu aces: a e iew o
undamen als and applica ions. Rep. P og. Phys. 81, 026401 (2018).
2. Shal ou , A. M., Shalae , V. M. & B onge sma, M. L. Spa io empo al ligh
con ol wi h ac i e me asu aces. Science 364, eaa 3100 (2019).
3. Wang, Q. e al. Op ically econfigu able me asu aces and pho onic de ices
based on phase change ma e ials. Na . Pho on 10,60–65 (2016).
4. Li, S. Q. e al. Phase-only ansmissi e spa ial ligh modula o based on
unable dielec ic me asu ace. Science 364, 1087–1090 (2019).
5. Wu, P. C. e al. Dynamic beam s ee ing wi h all-dielec ic elec o-op ic III–V
mul iple-quan um-well me asu aces. Na . Commun. 10, 3654 (2019).
6. Hols een, A. L., Cihan, A. F. & B onge sma, M. L. Tempo al colo mixing and
dynamic beam shaping wi h silicon me asu aces. Science 365, 257–260
(2019).
7. an de G oep, J. e al. Exci on esonance uning o an a omically hin lens.
Na . Pho on 14, 426–430 (2020).
8. Shalagino , M. Y. e al. Reconfigu able all-dielec ic me alens wi h di ac ion-
limi ed pe o mance. Na . Commun. 12, 1225 (2021).
9. Zhang, Y. e al. Elec ically econfigu able non- ola ile me asu ace using low-
loss op ical phase-change ma e ial. Na . Nano echnol. 16, 661–666 (2021).
10. Wang, Y. e al. Elec ical uning o phase-change an ennas and me asu aces.
Na . Nano echnol. 16, 667–672 (2021).
11. Yu, P., Li, J. & Liu, N. Elec ically unable op ical me asu aces o dynamic
pola iza ion con e sion. Nano Le . 21, 6690–6695 (2021).
12. Biswas, S., G ajowe , M. Y., Wa anabe, K., Taniguchi, T. & A wa e , H. A.
B oadband elec o-op ic pola iza ion con e sion wi h a omically hin black
phospho us. Science 374, 448–453 (2021).
13. Nicholls, L. H. e al. Ul a as syn hesis and swi ching o ligh pola iza ion in
nonlinea aniso opic me ama e ials. Na . Pho on 11, 628–633 (2017).
14. Schi a o, A. e al. T ansien op ical symme y b eaking o ul a as b oadband
dich oism in plasmonic me asu aces. Na . Pho on 14, 723–727 (2020).
15. Wang, K. e al. High con as , em osecond ligh pola iza ion manipula ion in
epsilon-nea -ze o ma e ial coupled o a plasmonic nanoan enna a ay. ACS
Pho onics 8, 2791–2799 (2021).
16. Zano o, S. e al. Op omechanics o chi al dielec ic me asu aces. Ad . Op .
Ma e . 8, 1901507 (2020).
17. Shi, Z. e al. Con inuous angle- unable bi e ingence wi h ee o m
me asu aces o a bi a y pola iza ion con e sion. Sci. Ad . 6, eaba3367
(2020).
18. Zhang, X. e al. Con olling angula dispe sions in op ical me asu aces. Ligh
Sci. Appl. 9, 76 (2020).
19. Chen, H.-T. In e e ence heo y o me ama e ial pe ec abso be s. Op .
Exp ess 20, 7165 (2012).
20. Ameling, R. & Giessen, H. Mic oca i y plasmonics: s ong coupling o
pho onic ca i ies and plasmons. Lase Pho on. Re . 7, 141–169 (2013).
21. Be khou , A. & Koende ink, A. F. Pe ec abso p ion and phase singula i ies in
plasmon an enna a ay e alons. ACS Pho onics 6, 2917–2925 (2019).
22. Meng, C. e al. Dynamic piezoelec ic MEMS-based op ical me asu aces. Sci.
Ad . 7, eabg5639 (2021).
23. Golds ein, D. H. Pola ized Ligh (Taylo and F ancis, 2010).
24. Wang, T., Sawada, R. & Lee, C. A piezoelec ic mic omachined ul asonic
ansduce using pis on-like memb ane mo ion. IEEE Elec on De ice Le . 36,
957–959 (2015).
25. Be khou , A., Wol e ink, T. A. W. & Koende ink, A. F. S ong coupling o
gene a e complex bi e ingence: me asu ace in he middle e alons. ACS
Pho onics 7, 2799–2806 (2020).
26. Dai, D., Bau e s, J. & Bowe s, J. E. Passi e echnologies o u u e la ge-scale
pho onic in eg a ed ci cui s on silicon: Pola iza ion handling, ligh non-
ecip oci y and loss educ ion. Ligh Sci. Appl. 1, e1 (2012).
27. He, C. e al. Pola isa ion op ics o biomedical and clinical applica ions: a
e iew. Ligh Sci. Appl. 10, 194 (2021).
28. Johnson, P. B. & Ch is y, R. W. Op ical cons an s o he noble me als. Phys.
Re . B 6, 4370–4379 (1972).
Acknowledgemen s
This esea ch has ecei ed unding om he VKR Founda ion (Awa d in Technical and
Na u al Sciences 2019, S.I.B. and G an No. 37372, F.D.); he EU Ho izon 2020 esea ch
and inno a ion p og am (Ma ie Skłodowska-Cu ie g an ag eemen No. 713694, C.M.);
as well as om he Resea ch Council o No way (P ojec numbe 323322, P.C.V.T.). C.M.
acknowledges Yao Xiao o he help in figu e p epa a ion, Ying Qu and Ma in Tho-
maschewski o hei help in he expe imen s. P.T. acknowledges Jon Vedum o helping
wi h he con ol elec onics o he MEMS mi o s.
Au ho con ibu ions
P.C.V.T. and C.M. pe o med he simula ions and designed he OMS, ab ica ed and
assembled he MEMS-OMS-based DWP samples. C.M. cons uc ed he expe imen al
se up, pe o med he measu emen s, and analysed he da a. All au ho s con ibu ed o
he p ojec idea, discussion o he esul s ob ained and w i ing he manusc ip . S.I.B.
supe ised he p ojec .
Compe ing in e es s
The pape au ho s along wi h Jo Gjessing and Ch is ophe Di dal om SINTEF a e
in en o s on a ela ed pa en applica ion filed by he Uni e si y o Sou he n Denma k
and SINTEF unde Uni ed Kingdom Pa en Applica ion No. 2113182.6.
Addi ional in o ma ion
Supplemen a y in o ma ion The online e sion con ains supplemen a y ma e ial
a ailable a h ps://doi.o g/10.1038/s41467-022-29798-0.
Co espondence and eques s o ma e ials should be add essed o Fei Ding o Se gey I.
Bozhe olnyi.
Pee e iew in o ma ion Na u e Communica ions hanks he o he anonymous
e iewe (s) o hei con ibu ion o he pee e iew o his wo k. Pee e iew epo s a e
a ailable
Rep in s and pe mission in o ma ion is a ailable a h p://www.na u e.com/ ep in s
Publishe ’s no e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in
published maps and ins i u ional a filia ions.
Open Access This a icle is licensed unde a C ea i e Commons
A ibu ion 4.0 In e na ional License, which pe mi s use, sha ing,
adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as you gi e
app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e
Commons license, and indica e i changes we e made. The images o o he hi d pa y
ma e ial in his a icle a e included in he a icle’s C ea i e Commons license, unless
indica ed o he wise in a c edi line o he ma e ial. I ma e ial is no included in he
a icle’s C ea i e Commons license and you in ended use is no pe mi ed by s a u o y
egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om
he copy igh holde . To iew a copy o his license, isi h p://c ea i ecommons.o g/
licenses/by/4.0/.
© The Au ho (s) 2022
NATURE COMMUNICATIONS | h ps://doi.o g/10.1038/s41467-022-29798-0 ARTICLE
NATURE COMMUNICATIONS | (2022) 13:2071 | h ps://doi.o g/10.1038/s41467-022-29798-0 | www.na u e.com/na u ecommunica ions 7
74 CHAPTER 5. ARTICLES
5.3 MEMS Tunable Me asu aces Based on Gap
Plasmon o Fab y–P´e o Resonances
Th ane P.C.V., Meng C., Ding F. and Bozhe olnyi S.I.
MEMS Tunable Me asu aces Based on Gap Plasmon o Fab y–P´e o Reso-
nances.
Nano Le e s (2022).
MEMS Tunable Me asu aces Based on Gap Plasmon o Fab y−Pe o
Resonances
Paul C. V. Th ane,
§
Chao Meng,
§
Fei Ding, and Se gey I. Bozhe olnyi*
Ci e This: h ps://doi.o g/10.1021/acs.nanole .2c01692
Read Online
ACCESS Me ics & Mo e A icle Recommenda ions *
sı Suppo ing In o ma ion
ABSTRACT: Tunable me asu aces p omise o enable adap i e
op ical sys ems wi h complex unc ionali ies. Among possible
ealiza ions, a ecen pla o m combining mic oelec omechanical
sys ems (MEMS) wi h gap-su ace plasmon (GSP) me asu aces
o e s high modula ion e iciency, b oadband ope a ion, and as
esponse. We compa e unable me asu aces ope a ing in GSP and
Fab y−Pe o (FP) egions by in es iga ing pola iza ion-independ-
en blazed g a ings bo h nume ically and expe imen ally. Peak
e iciency is calcula ed o be ∼75% in bo h cases (∼40% in
measu emen s), while he ope a ion bandwid h is ound la ge
when ope a ing in he GSP egion. Ad an ages o ope a ing in he
FP egion include elaxed assembly equi emen s and ope a ion
ole ances. Addi ionally, simula ion and expe imen al esul s show ha coupling be ween neighbo ing uni cells inc eases o la ge
ai gaps, esul ing in de e io a ed e iciency. We belie e he p esen ed analysis p o ides impo an guidelines o designing unable
me asu aces o di e se applica ions in minia u ized adap i e op ical sys ems.
KEYWORDS: Me asu ace, Tunable, MEMS, Gap Su ace Plasmon, Plasmonic, Fab y−Pe o , In e cell Coupling
■INTRODUCTION
Me asu aces ha e success ully demons a ed a wide ange o
op ical e ec s and componen s,
1−5
wi h a lo o ecen esea ch
ocusing on de eloping me asu aces wi h unable p ope ies o
enable adap i e op ical componen s and wi h se e al di e en
echniques being ollowed, each ha ing hei own ad an ages
and disad an ages.
6−9
One me hod o achie e his unabili y is,
o ins ance, o include ma e ials ha unde go a phase change.
GeSbTe can, o example, change om ha ing a c ys al
s uc u e o an amo phous s a e depending on he empe -
a u e, wi h he wo s a es ha ing e y di e en pe mi i i y.
10
By inco po a ing esis i e hea e s, i is hus possible o make
me asu ace elemen s ha can change hei esonances qui e
signi ican ly wi h a d awback being ha demons a ed de ices
ha e slow swi ching ime.
11
Fas e esponses ha e been
demons a ed using he elec o-op ic e ec in li hium
nioba e
12,13
o by modula ing he ee ca ie densi y using
elec ic
14
o op ical
15
signals, wi h an issue being ha he
pe mi i i y changes a e limi ed o hin accumula ion o
deple ion laye s gi ing low modula ion anges.
7
The e ec can
be enhanced by using ε-nea -ze o ma e ials
16
o by using 2D
ma e ials such as g aphene
17
o black phospho us.
18
Liquid
c ys als enable la ge and mo e e icien modula ion by
changing he e ac i e index a ound he nanos uc u es
7
bu
again ha e slowe esponses due o he ime i akes o o a e
he molecules.
19
Me asu aces can also be adjus ed h ough
mechanically al e ing he sys em, wi h demons a ed concep s
including embedding he nanos uc u es in a s e chable
polyme
20
o inco po a ing he me asu ace wi h MEMS.
21
MEMS based unable me asu aces can achie e high e iciency
modula ion while s ill swi ching as enough o many
applica ions depending on he speci ic mechanical implemen-
a ion, wi h mos sys ems being able o ope a e in he ange
om one kHz up o se e al hund ed kHz.
22
Fo isible and
nea -IR equencies he indi idual me a-a oms a e so small ha
indi idual ac ua ion by MEMS is challenging, while collec i e
modula ion o all me a-a oms is mo e s aigh o wa d.
One such ecen ly demons a ed pla o m
23
consis s o a
gold MEMS mi o
24
and a glass subs a e wi h gold
nanos uc u es, whe e he ai gap be ween he nanos uc u es
and mi o can be con olled accu a ely. The sys em is
designed o unc ion as a e lec i e op ical me asu ace (OMS)
o ligh wi h wa eleng h λ= 800 nm when he ai gap is less
han 50 nm. Fo hese small sepa a ions he e a e GSP
esonances
25
due o he nea ield coupling o he
nanos uc u es and mi o . By mo ing he mi o away, hese
GSP esonances disappea , swi ching o he me asu ace
Recei ed: Ap il 27, 2022
Re ised: Augus 12, 2022
Le e pubs.acs.o g/NanoLe
© XXXX The Au ho s. Published by
Ame ican Chemical Socie y A
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
Downloaded ia SINTEF on Augus 18, 2022 a 17:07:21 (UTC).
See h ps://pubs.acs.o g/sha ingguidelines o op ions on how o legi ima ely sha e published a icles.
unc ionali y and eplacing i wi h ha o a s anda d mi o .
The expe imen ally demons a ed e iciency o his sys em was
50%, wi h swi ching imes less han 0.4 ms. In his wo k we
desc ibe how he same MEMS-OMS pla o m can unc ion
also o la ge ai gaps owing o hyb id plasmonic FP
esonances.
26,27
This con igu a ion has ecen ly been used o
achie e e icien and as 0−2πbi e ingence con ol in
e lec ion.
28
No only is ab ica ion easie a la ge ai gaps
since any pa icle o une enness may obs uc he MEMS
mi o om ge ing close enough o he GSP esonances, bu
la ge gaps could also help educe he ade-o be ween
ape u e size and swi ching speed by alle ia ing squeeze ilm ai
damping in he sys em.
29
Addi ionally, he amoun o
simula ions equi ed o design is educed h ough he use o
he analy ic FP equa ion, emo ing he need o simula e he
esponse o e e y ai gap sepa a ely. We show also ha he
simula ed peak e iciency is a ound 75% o me asu aces
wo king in bo h GSP and FP egions, while he bandwid h is
la ge o he GSP me asu ace wi h a ound 2 imes he
bandwid h when compa ing wi h he me asu ace wo king a
he i s FP esonance. Fo la ge ai gaps he e is p og essi ely
mo e coupling/c oss- alk be ween neighbo ing nanos uc u es
due o sca e ing and mul iple e lec ions in he FP ca i y,
esul ing in a g adual dec ease in me asu ace e iciency. This is
a esul o he me asu ace design being based on simula ions
whe e he uni cells a e placed in a ays o iden ical s uc u es,
whe eas he ac ual me asu ace may gene ally consis o a ying
me a-a oms. We e i y his e ec bo h nume ically and
expe imen ally.
■RESULTS AND DISCUSSION
To compa e plasmonic me asu aces designed o wo k in he
GSP and FP egions, we i s calcula e he complex e lec ion
coe icien o di e en nanos uc u e geome ies a wo
nanos uc u e−subs a e sepa a ions co esponding o he
wo egions. Figu e 1a and Figu e 1b illus a e he MEMS-
OMS and i s cons i uen uni cells used in his wo k.
Speci ically, he pe iodically epea ed uni cell has a side
leng h Λand consis s o a gold nanob ick wi h hickness mand
side leng hs Lxand Lyand sepa a ed om a gold subs a e by
an ai gap Ta. In physical implemen a ions he e needs o be a
dielec ic subs a e suppo ing he nanob icks; his has been
omi ed in he simula ions excep when compa ing wi h he
expe imen al measu emen s discussed la e . To design he
dynamic MEMS-OMS, we se he wo king wa eleng h a λ=
800 nm and choose he uni cell size o 250 nm o a oid any
high-o de di ac ion and exci a ion o su ace wa es. Mean-
while, he op imal nanob ick hickness mis ound o be 50 nm,
ensu ing la ge e lec ion ampli udes and wide phase co e -
age.
23,30
Figu e 1c and Figu e 1d show elec ic ield plo s o
wo con igu a ions o he uni cell, while Figu e 1e and Figu e
1 show he e lec ion coe icien s as a unc ion o Lxand Ly o
Figu e 1. MEMS-OMS uni cell design wi hin espec i e GSP and FP egions. (a, b) Schema ic illus a ion o he MEMS-OMS me asu ace and
uni cell. A gold b ick wi h side leng hs Lxand Lyand hickness mo 50 nm is si ua ed a dis ance Taaway om a gold subs a e. The uni cell has a
squa e oo p in wi h side leng hs Λ= 250 nm. (c, d) The no m o he elec ic ield in he xz plane a he cen e o he nanob icks o x-pola ized
exci a ion a no mal incidence wi h sepa a ion dis ances o Ta= 20 nm and Ta= 430 nm, espec i ely. The nanob ick geome ies a e indica ed as
black squa es in (e) and ( ). (e, ) Absolu e alue o he complex e lec ion coe icien s calcula ed as a unc ion o nanob ick dimensions Lxand Ly
a he wa eleng h o λ= 800 nm o (e) Ta= 20 nm and ( ) Ta= 430 nm. The colo maps ep esen he e lec ion ampli ude o x-pola ized
exci a ion a no mal incidence, while he blue and black con ou lines indica e he e lec ion phases acqui ed o x- and y-pola ized exci a ions,
espec i ely. Blue ci cles indica e nanob ick geome ies o composing he pola iza ion independen MEMS-OMS blazed g a ings op imized in
espec i e GSP and FP egions. The phase and e lec ion ampli udes o hese nanob icks a e shown in Figu e S1.
Nano Le e s pubs.acs.o g/NanoLe Le e
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
B
wo di e en alues o Ta o no mally inciden x-pola ized
exci a ion a λ= 800 nm. A Ta= 20 nm, he e is a GSP
esonance a ound Lx= 150 nm wi h nea ield coupling
be ween he subs a e and nanob ick as can be seen in he plo
o he elec ic ield in Figu e 1c. A Ta= 430 nm, he e is a less
sha p esonance cen e ed a ound Lx= 160 nm, due o a hyb id
plasmonic/FP esonance be ween he subs a e and he laye
o nanob icks since he nea - ield coupling be ween he
nanob icks and subs a e is negligible, as shown in Figu e 1d.
As can be seen by compa ing he phase con ou s in Figu e 1e
and Figu e 1 , he ange o a ailable e lec ion phases a e
sligh ly la ge o he case whe e Ta= 20 nm; howe e he
di e ence is no signi ican enough o gi e la ge e iciencies
o he blazed g a ing designs p esen ed la e . The phase
p o iles o hese g a ings a e shown in Figu e S1.
The ansi ion be ween hese GSP and FP egions is
displayed in Figu e 2, whe e he e lec ion coe icien s o
di e en b ick sizes a e shown as a unc ion o Ta. The i s FP
esonance is loca ed a ound Ta= 350 nm wi h Ta+Tm/2 close
o λ/2, wi h a di e ence co esponding o he phase change
upon e lec ion on he gold mi o . A his sepa a ion he
sys em ac s as a gold mi o wi h he e lec ion coe icien
being independen o Lxand Ly, since a his sepa a ion
dis ance he nanos uc u es a e cen e ed in he in e e ence
minimum om he supe posi ion o inciden and elec ed
ields. Fo sligh ly la ge ai gaps he e lec ion is e y
dependen on he nanob ick dimensions (e.g., Ta= 430 nm
Figu e 2. Complex e lec ion coe icien s as a unc ion o ai gap and nanob ick dimensions. The colo map ep esen s he e lec ion ampli udes o
no mally inciden x-pola ized ligh , while he con ou lines indica e he e lec ion phases. The h ee sub igu es ep esen h ee di e en cases o he
nanob ick dimensions, namely, (a) squa e b icks (Lx=Ly), (b) cons an Lx= 100 nm, and (c) cons an Ly= 100 nm. The ligh is no mally inciden
and wi h wa eleng h 800 nm.
Figu e 3. Dynamic MEMS-OMS blazed g a ings designed o espec i e GSP and FP egions. (a) Supe cell ske ches o he 12-elemen
pola iza ion-independen dynamic MEMS-OMS blazed g a ings. Ou lines o he nanob ick dimensions op imized o Ta= 20 nm and Ta= 430 nm
a e indica ed wi h black squa es and o ange dashed lines, espec i ely. (b, c) Calcula ed di ac ion e iciencies in o he specula (m= 0) and i s
di ac ion o de (m= +1) as a unc ion o wa eleng h wi h he op imal ai gap o each g a ing, o x- and y-pola ized exci a ions. (d, e) Calcula ed
di ac ion e iciencies as a unc ion o ai gap Taa λ= 800 nm o x- and y-pola ized exci a ions.
Nano Le e s pubs.acs.o g/NanoLe Le e
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
C
as shown in Figu e 1 ) and hen g adually e u ns o mi o -like
beha io a he second FP esonance loca ed a ound Ta+Tm/2
close o λ, wi h he pa e n epea ing o subsequen FP apa
by a spaced sepa a ion o λ/2, again wi h a small co ec ion
due o he phase change upon mul iple e lec ions on he gold
mi o . This can be accu a ely desc ibed wi h he FP
equa ion:
26
e
1 e
i kn T
i kn T
o 12
12 21 23
2 cos( )
21 23
2 cos( )
2 a 2
2 a 2
= +
(1)
whe e he o al e lec ion coe icien o is gi en as a unc ion
o he e lec ion and ansmission coe icien s ij and ij, wi h
he ligh inciden on egion j om egion iand he subsc ip s 1,
2, and 3 espec i ely e e ing o he egions abo e he
nanob icks, be ween he nanob icks and subs a e, and he
subs a e. No e ha he e lec ion and ansmission coe icien s
in gene al a e dependen on he pola iza ion and incidence
angle o he ligh . Wi h mi o -symme ic nanos uc u es and
bo h egions 1 and 2 consis ing o he same ma e ial (e.g., ai ),
we ha e 12 = 21, 12 = 21, and n2= 1. Tacos(θ2) is he e ec i e
ai gap o ligh a e sing he gap wi h an angle θ2and can be
simpli ied as Ta o no mal incidence. By simula ing he
s uc u es wi hou any gold subs a e, he e lec ion and
ansmission coe icien s ij and ij a e de e mined o each
nanob ick geome y. Equa ion 1 is hen used o calcula e he
o al e lec ion coe icien as a unc ion o Ta, hus a oiding he
equi emen o simula ing he ull s uc u e o e e y ai gap
sepa a ion. This me hod gi es co ec esul s excep o e y
small ai gaps, in his case Ta< 80 nm = λ/10, whe e nea - ield
Figu e 4. E ec o coupling ia mi o subs a e on g a ing e iciency: compa ison o simula ions and expe imen al measu emen s. (a) Supe cell o a
MEMS-OMS dynamic blazed g a ing wi h 12 me a-a oms op imized o he GSP egion. (b) Di ac ion e iciencies as a unc ion o ai gap Ta. Fo
hese simula ions he gold nanob icks a e placed on a glass subs a e, which is he case o he ab ica ed OMS shown in (c). The OMS is placed in
close p oximi y (<3 μm) o a piezoelec ic MEMS gold mi o . The ai gap can hen be changed by applying ol ages on he MEMS. The measu ed
di ac ion e iciencies a e shown in (d). The ai gap alues in (d) ha e been added by measu ing he app oxima e ela ionship be ween ai gap and
ol age as desc ibed in Figu e S7. In bo h (b) and (d) he uppe (lowe ) plo is o x-pola ized (y-pola ized) exci a ion. The nanob ick hickness is
50 nm, and he wa eleng h is 800 nm.
Nano Le e s pubs.acs.o g/NanoLe Le e
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
D
coupling and co esponding GSP exci a ion mus be aken in o
accoun . The di e ence be ween he esul s om eq 1 and ull-
wa e simula ions including he subs a e can be clea ly
obse ed o small gap sizes in Figu e S2.
A e analyzing he pola iza ion-dependen esponses o he
uni cell in bo h GSP and FP egions, we s a o implemen
unc ional me asu aces. Figu e 3 compa es he pe o mance o
wo blazed me asu ace g a ings, one op imized o Ta= 20 nm
and he o he o Ta= 430 nm. As illus a ed in Figu e 3a, he
me asu ace g a ings consis o a pe iodic a ay o 12 elemen s,
wi h he i s 2 elemen s emp y while he o he dimensions a e
chosen and ma ked wi h blue ci cles in Figu e 1e and Figu e 1
o p o ide la ge e lec ion ampli udes and an app oxima ely
linea phase g adien along he x-di ec ion.
In gene al he me asu ace can be designed o con ol wo
o hogonal pola iza ion s a es independen ly by using aniso-
opic elemen s, bu he blazed g a ing is made pola iza ion
independen by choosing iso opic elemen s wi h Lx=Ly.
Figu e 3b and Figu e 3c show he di ac ion e iciencies o he
wo g a ings o x- and y-pola ized ligh as a unc ion o
wa eleng h, whe e he pola iza ion independen beha io can
be obse ed. The e iciencies o he +1 di ac ion o de a he
design wa eleng h o λ= 800 nm a e simila o bo h g a ings,
wi h he alues app oaching 75%. Howe e , he ope a ing
bandwid h is di e en . The g a ing wo king a Ta= 20 nm has
an e iciency abo e 60% in he wa eleng h ange be ween 760
and 900 nm, while o he g a ing wo king a Ta= 430 nm he
co esponding wa eleng h ange is only spanning om 770 o
825 nm. This educed bandwid h is due o he ac ha he FP
esonance is na owe han he GSP esonance. A highe o de
FP esonances, he bandwid h will be educed e en u he as
he esonance equi es he ai gap o be an in ege mul iple o
hal wa eleng hs. Con e sely, his e ec migh be used o
design highly ch oma ic me asu aces by choosing a la ge ai
gap. Figu e 3d and Figu e 3e compa e he di ac ion
e iciencies o wo me asu ace g a ings o x- and y-pola ized
ligh as a unc ion o Taa he design wa eleng h o 800 nm.
Imp essi ely, bo h me asu ace g a ings achie e mo e han 65%
e lec ion in he +1 di ac ion o de when he ai gap is Ta=
20 nm and Ta= 430 nm due o he simila i y o he me a-
a oms. Bu he e is a clea gain in e iciency by ailo ing he
me a-a oms o he ele an ai gap, which is ue o bo h
pola iza ion s a es.
One may expec he same esponses o me asu ace blazed
g a ings wi h epea ing FP egions. Howe e , in ou
simula ions and expe imen al esul s we obse e a dec ease
in he +1 o de e iciency a highe o de FP esonances in
combina ion wi h mo e e lec ion in o o he di ac ion o de s,
which can be unde s ood as inc eased coupling o c oss- alk
be ween neighbo ing nanos uc u es ia e lec ions in he gold
subs a e. Figu e 4a shows ano he 12-elemen blazed g a ing
op imized o Ta= 20 nm. As ea lie , he choice o nanob ick
dimensions is based on simula ions whe e he nanob icks a e
placed in an a ay o iden ical neighbo s, while in p ac ical
applica ions he neighbo s may ha e any geome y. This has
been shown o no signi ican ly a ec he pe o mance o GSP
me asu aces as long as he phase g adien is no oo la ge.
25
Howe e , as can be seen in he e lec ion ampli ude o he
di ac ion o de s plo ed in Figu e 4b, he e iciency o he
g a ing deg ades as Tainc eases, wi h ligh going in o
unwan ed di ac ion modes o he han he desi ed +1
di ac ion o de , alling om 72% a Ta= 430 nm o 66% a
he ou h FP esonance due o inc eased coupling be ween
elemen s wi hin he supe cell ia he mi o subs a e. In
Figu e S3, he same e ec is isible o a g a ing made o 8 uni
cells. Wi h ewe me a-a oms he neighbo ing nanob icks a e
less simila in size, esul ing in a la ge d op in e iciency going
om a ound 70% a Ta= 430 nm o less han 50% a he ou h
FP esonance. The e lec ed ield dis ibu ions o a blazed
g a ing a se e al di e en ai gap sepa a ions a e shown in
Figu e S4.Figu e 4c and Figu e 4d show expe imen al
measu emen s o a ab ica ed me asu ace pai ed wi h a
piezoelec ic MEMS mi o , showing he same g adual
dec ease o he maximum di ac ion e iciency om a ound
38% a he i s FP esonance o 30% a he ou h FP
esonance o bo h pola iza ion s a es, oge he wi h inc eased
in ensi y in he +2 di ac ion o de . This coupling issue is
especially impo an when making high NA lenses o o he
componen s equi ing la ge de lec ion angles, whe e he la ge
phase g adien will equi e nanob icks ha di e signi ican ly
om hei neighbo s. De ails and some discussion o he
ab ica ion and op ical cha ac e iza ion can be ound in Figu es
S5 and S6, while Figu e S7 desc ibes he measu emen s done
o de e mine he ela ionship be ween ai gap and ol age
applied o he MEMS mi o . I should be no ed ha he
minimal ai gap achie ed wi h he measu ed sample was ∼150
nm, su icien o FP ope a ion bu no op imal o GSP
ope a ion. Close sepa a ions can be achie ed
23
bu likely
equi es signi ican ly mo e e o o be p oduced wi h high
yield and migh be ha de o ealize wi h la ge ape u es. To
conclude, we show how he ecen ly de eloped MEMS-OMS
pla o m is no limi ed o wo king in he GSP egion. Fo
la ge ai gaps he FP esonances enable he sys em o s ill
unc ion as a me asu ace, wi h sligh ly smalle phase ange and
simila e iciencies i he nanos uc u e geome ies a e
op imized o wo k a he ele an ai gap. The main ad an age
o allowing o la ge ai gaps is alle ia ing issues wi h hin ilm
damping o high speed ope a ion, as well as simpli ying
ab ica ion ole ances as dec easing he ai gap below 50 nm is
a e y challenging p oblem, equi ing la pa allel su aces ee
om any pa icles o i egula i ies ha may obs uc he
MEMS mo emen . Meanwhile, wo king in he GSP egion
gi es be e bandwid h and ewe issues wi h coupling be ween
me a-a oms, which causes he sys em o change beha io
be ween di e en FP pe iods when ha ing me asu aces
comp ised o noniden ical me a-a oms. Ul ima ely, he choice
be ween wo di e en , albei simila , ope a ion egimes o he
conside ed MEMS-OMS pla o m should be made by ca e ully
conside ing all implica ions o hei ad an ages and d awbacks,
highligh ed in his wo k, o a ge ed unc ionali ies and
pa icula applica ions in op ical sys ems.
■METHODS
The simula ions we e done using COMSOL Mul iphysics 5.6
wi h he Wa e Op ics module. The e ac i e index o gold
was in e pola ed om expe imen al alues
31
o bo h he gold
subs a e and gold nanob icks. When simula ing indi idual
nanob icks, he uni cell is se o ha e pe iodic condi ions in
bo h x- and y-di ec ions, while he gold subs a e is backed by a
pe ec elec ical conduc o condi ion and he ai egion is
padded wi h a pe ec ly ma ched laye . When using eq 1 he
coe icien s 12, 21, 12, and 21 a e de e mined by simula ing he
nanob icks wi hou he gold subs a e, wi h pe ec ly ma ched
laye s backing he domains on bo h sides o he nanos uc u e
in he z-di ec ion. 23, he e lec ion coe icien o he gold
subs a e o ligh wi h no mal incidence, is de e mined by
Nano Le e s pubs.acs.o g/NanoLe Le e
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
E
n n
n n
23
2 3
2 3
=
+
n2being he e ac i e index o he laye abo e he subs a e (in
his case ai , n2= 1) and n3 he e ac i e index o gold. The
ligh was se o be no mally inciden o all simula ions, and he
esul s we e calcula ed independen ly o ligh linea ly
pola ized in he x- and y-di ec ions. When simula ing g a ings
wi h la ge ai gaps and many nanob icks, he simula ion
olume was educed by simula ing hal he uni cell and
eplacing he pe iodic bounda y condi ions in he y-di ec ion
by pe ec ly magne ic o elec ic conduc o s depending on he
inciden pola iza ion s a e. Fo he simula ions in Figu e 4, he
nanob icks we e placed on a lossless dielec ic ma e ial wi h
e ac i e index 1.46.
Fab ica ion o he MEMS me asu ace de ices is done by
indi idually manu ac u ing MEMS mic omi o s and glass
subs a es con aining he me asu aces, be o e manually gluing
he MEMS chips and glass subs a es oge he in a clean oom
en i onmen . A desc ip ion and discussion o his p ocess can
be ound in Figu e S5. Addi ional de ails can also be ound in
e s 24 and 32 o MEMS ab ica ion and e s 23 and 28 o he
MEMS−me asu ace combina ion.
Op ical cha ac e iza ion is done by sending lase ligh wi h
wa eleng h 800 nm h ough a linea pola ize , hal wa e pla e
(used o swi ch be ween wo linea pola iza ion s a es), and
beam spli e and hen ocusing on o he sample using a
mic oscope objec i e. The ligh is e lec ed back in o he
objec i e and is edi ec ed by he beam spli e , be o e ube
lens, i is (in image plane o spa ial il e ing), and wo lenses
ha elay he ligh on o a CMOS came a. The las lens can be
lipped in and ou o he op ical pa h o swi ch be ween
cap u ing he di ec and Fou ie images. The di ec objec
image is used o ensu e he signal is collec ed om only he
me asu ace a ea, while he in ensi y in he di e en di ac ion
o de s is measu ed by in eg a ing he in ensi y o he
co esponding a eas in he Fou ie plane image. De ails on
he equipmen and a diag am o he se up can be ound in
Figu e S6.
■ASSOCIATED CONTENT
*
sı Suppo ing In o ma ion
The Suppo ing In o ma ion is a ailable ee o cha ge a
h ps://pubs.acs.o g/doi/10.1021/acs.nanole .2c01692.
Figu e S1 showing phase and ampli ude p o iles o he
blazed me ag a ings; Figu e S2 showing e o o eq 1;
Figu e S3 showing simula ion esul s o a me asu ace
blazed g a ing wi h 8 elemen s; Figu e S4 showing plo s
o e lec ed ield dis ibu ion o a blazed g a ing a
se e al ai gap alues; Figu e S5 showing de ails o he
ab ica ion o he MEMS me asu ace de ices wi h
ele an discussion; Figu e S6 showing a schema ic o
op ical measu emen se up; Figu e S7 showing cha ac-
e iza ion and discussion o ai gap as a unc ion o
ol age (PDF)
■AUTHOR INFORMATION
Co esponding Au ho
Se gey I. Bozhe olnyi −Cen e o Nano Op ics, Uni e si y o
Sou he n Denma k, Odense DK-5230, Denma k;
o cid.o g/0000-0002-0393-4859; Email: seib@
mci.sdu.dk
Au ho s
Paul C. V. Th ane −Cen e o Nano Op ics, Uni e si y o
Sou he n Denma k, Odense DK-5230, Denma k; SINTEF
Sma Senso s and Mic osys ems, 0737 Oslo, No way;
o cid.o g/0000-0001-5296-2912
Chao Meng −Cen e o Nano Op ics, Uni e si y o Sou he n
Denma k, Odense DK-5230, Denma k; o cid.o g/0000-
0002-2126-6954
Fei Ding −Cen e o Nano Op ics, Uni e si y o Sou he n
Denma k, Odense DK-5230, Denma k; o cid.o g/0000-
0001-7362-519X
Comple e con ac in o ma ion is a ailable a :
h ps://pubs.acs.o g/10.1021/acs.nanole .2c01692
Au ho Con ibu ions
§
P.C.V.T. and C.M. con ibu ed equally o his wo k.
No es
The au ho s decla e he ollowing compe ing inancial
in e es (s): The pape au ho s along wi h J. Gjessing and C.
Di dal om SINTEF a e in en o s on a ela ed pa en
applica ion led by he Uni e si y o Sou he n Denma k and
SINTEF unde Uni ed S a es Pa en Applica ion No. 17/
467542.
■ACKNOWLEDGMENTS
The au ho s hank C. Di dal o use ul discussions and
eedback on he manusc ip . This esea ch is suppo ed by he
Resea ch Council o No way (P ojec 323322), he VKR
Founda ion (Awa d in Technical and Na u al Sciences 2019
and G an 37372), and he EU Ho izon 2020 Resea ch and
Inno a ion P og amme (Ma ie Skłodowska-Cu ie G an
Ag eemen 713694).
■REFERENCES
(1) Qiu, C. W.; Zhang, T.; Hu, G.; Ki sha , Y. Quo Vadis,
Me asu aces? Nano Le . 2021,21, 5461−5474.
(2) Scheue , J. Op ical Me asu aces A e Coming o Age: Sho - And
Long-Te m Oppo uni ies o Comme cial Applica ions. ACS
Pho onics 2020,7, 1323−1354.
(3) Chen, W. T.; Zhu, A. Y.; Capasso, F. Fla Op ics wi h
Dispe sion-Enginee ed Me asu aces. Na u e Re iews Ma e ials 2020,
5, 604−620.
(4) Kamali, S. M.; A babi, E.; A babi, A.; Fa aon, A. A Re iew o
Dielec ic Op ical Me asu aces o Wa e on Con ol. Nanopho onics
2018,7, 1041−1068.
(5) Ding, F.; Po s, A.; Bozhe olnyi, S. I. G adien me asu aces: a
e iew o undamen als and applica ions. Rep. P og. Phys. 2018,81,
026401.
(6) Hail, C. U.; Michel, A. K. U.; Poulikakos, D.; Eghlidi, H. Op ical
Me asu aces: E ol ing om Passi e o Adap i e. Ad anced Op ical
Ma e ials 2019,7, 1801786.
(7) Shal ou , A. M.; Shalae , V. M.; B onge sma, M. L.
Spa io empo al ligh con ol wi h ac i e me asu aces. Science 2019,
364, eaa 3100.
(8) Yang, J.; Gu ung, S.; Bej, S.; Ni, P.; Lee, H. W. H. Ac i e Op ical
Me asu aces: Comp ehensi e Re iew on Physics, Mechanisms, and
P ospec i e Applica ions. Rep. P og. Phys. 2022,85, 036101.
(9) Du, K.; Ba kaoui, H.; Zhang, X.; Jin, L.; Song, Q.; Xiao, S.
Op ical me asu aces owa ds mul i unc ionali y and unabili y.
Nanopho onics 2022,11, 1761−1781.
(10) Ding, F.; Yang, Y.; Bozhe olnyi, S. I. Dynamic Me asu aces
Using Phase-Change Chalcogenides. Ad anced Op ical Ma e ials 2019,
7, 1801709.
Nano Le e s pubs.acs.o g/NanoLe Le e
h ps://doi.o g/10.1021/acs.nanole .2c01692
Nano Le . XXXX, XXX, XXX−XXX
F
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1560
02π
φ(x,y,Ta1)
0 2π
φ(x,y,Ta2)
Va1 Va2
(b)
x
y
TM
TE To al m=0 m=−1 m=+1 m=+2m=−2
( )
Inciden
Beam +1 O de −1 O de
VMEMS=Va1
VMEMS=Va2
BMS
MEMS
(a)
x(TM)
z
y(TE)
k
MEMS-BMS BDG
1
∠
| |
Supe cell o MEMS-BMS BDG
-400°
-200°
0°
200°
400°
(c)
0.9
0.8
0.7
0.6
0.5
Ta=150 nm+q2λ/2
Ta=350 nm+q1λ/2
23 4 567 8 910 11 12
12Λ
Λ
MS1
MS2
Λ
x
y
Ai Gap Ta (nm)
0.8
0.6
0.4
0.2
00 500 1000 1500 2000 2500 3000 3500 4000
Re lec ance
Ta = 350 nm Ex
(V/m)
1
0
-1
Ta = 150 nm
Re lec. Field (TM)
(d)
x
z
y
Eno m
(V/m)
0
1
2
3
4
5
6
Ta = 350 nm
Ta= 150 nm
To al Field (TM)
(e)
x
z
y
Fig. 3. MEMS-BMS BDG wi h econ igu able di ac ion o de s: design. (a) Schema ic ende ing o he MEMS-BMS BDG o ac i e e lec ed beam
swi ching be ween he +1 and −1 di ac ion o de s by elec ically ac ua ing he MEMS-BMS be ween wo dis inc ai gaps wi h di e en ol ages.
(b) Re e sible 1D linea phase g adien s o he supe cell a ai gaps o Ta1=(350 +q1×λ/2)nm and Ta2=(150 +q2×λ/2)nm. (c) Calcula ed e lec-
ion ampli udes (black ma ke s) and phases ( ed ma ke s) o 12 selec ed uni cells o cons uc ing he MEMS-BMS BDG supe cell, wi h an inciden
wa eleng h o λ=800 nm. (d), (e) Simula ed e lec ed (d) and o al (e) elec ic ields o he MEMS-BMS BDG supe cell unde a TM exci a ion wi h
λ=800 nm o Ta=150 and 350 nm, espec i ely. ( ) Calcula ed di ac ion e iciencies o di e en o de s (|m| ≤ 2) as a unc ion o he ai gap sizes Ta
o bo h TM and TE exci a ions wi h λ=800 nm.
o 1750 and 1950 nm (i.e., q1=q2=4, Fig. S13), he e icien-
cies s ill emain a ∼65% (m= −1) and ∼70% (m= +1) bu
dec ease sligh ly due o s onge in e cell coupling o inc eased
ai gap sizes [31,58]. O e all, he MEMS-BMS BDG designed
p omises a high-con as dynamic swi ching capabili y be ween
±1 di ac ion o de s wi hin a la ge MEMS-BMS sepa a ion ange
[Fig. 3( )]. Fu he mo e, i s ope a ion is obus wi h espec o
he ope a ion wa eleng h (Fig. S14), main aining a high quali y
when di ac ing in o +1 o de a Ta=q1×λ/2 o e a b oad
wa eleng h ange, al hough exhibi ing somewha s onge de e io-
a ion when di ac ing in o −1 o de . The la e is ela ed o he
ci cums ance ha he −1 o de di ac ionin ol es bo h MS laye s
and he in e ening SU-8 laye , making i sensi i e o he inciden
wa eleng h. In con as , he +1 o de di ac ion is go e ned solely
by he MS1 laye , ea u ing a b oad bandwid h ope a ion simila
o he MEMS- unable single-laye MS cases [20,31]. Finally, he
MEMS-BMS BDG ope a ion was also ound obus wi h espec
o he in luence o inc eased loss o he gold me a-a oms (Fig. S15),
mode a e a ia ions in SU-8 hickness (Fig. S16), and de ia ions in
me a-a om sizes (Figs. S17).
D. MEMS-BMS Blazed Di ac ion G a ing:
Cha ac e iza ion
The MEMS-BMS BDG designed abo e was assembled om
a sepa a ely ab ica ed BMS on a glass subs a e (14 mm×
14 mm ×400 µm), a as -speed ul a- la piezoelec ic MEMS
mi o (500 µm in diame e , Fig. S19), and a p in ed ci cui boa d
(see Appendix A and Supplemen 1 S12). The ab ica ed BMS
was inspec ed wi h op ical and elec on mic oscopy a e each
MS1, SU-8, and MS2 laye ab ica ion s ep o moni o me a-
a om geome ies and sizes, SU-8 hickness, and alignmen quali y
be ween he wo MS laye s [Fig. 4(a) and Figs. S21 o S23]. The
MEMS gold mi o s we e also checked wi h op ical mic oscopy o
a oid any con amina ions on he mi o su ace. A e assembling,
he MEMS-BMS sepa a ion was cha ac e ized by examining
he e lec ion spec a o he MEMS-BMS using a b oadband
ligh sou ce (Fig. S25). The ini ial ai gap dis ance was ∼2.5 µm,
de e mined mainly by he iscosi y and amoun o he glue we
applied du ing assembly. By ac ua ing he inne (o ou e ) MEMS
elec odes, he MEMS mi o can be mo ed close o (o away
om) he BMS. The o al ai gap (Ta) uning ange, achie ed by
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1561
(a)
0.5 μm
Glass
SU8
MS1
MS2
BMS
Glass
MS1
MS1
0.1
0.2
0.3
0.4
Di ac ion E iciency
Ai Gap Ta (nm)
0
0.5
1
-0.5
0180020002200240026002800300032003400 -1
Di ac ion Con as
5
0 10 15 20
2
5
10
15
20
VMEMS (V)
Inne Elec odes Ou e Elec odes
(d)
TM
TE m=-1 m=+1 Con as
q=4 q=5 q=6 q=7
(e)
Time (μs)
𝜏 ise=6 μs𝜏 all=5 μs
𝜏 ise=5 μs
𝜏 all=6 μs
Vm
-400 -300 -200 -100 0 100 200 300 400
0
10
20
30
20
30
9
11
VMEMS (V)
In ensi y (mV)
m=-1 m=+1
(b)
kx (k0)
Ai Gap Ta (nm)
0.50.40.3
0.20.10
-0.5-0.4-0.3-0.2-0.1
1800
2000
2200
2400
2600
2800
3000
3200
3400 λ=850 nm
0
0.2
0.4
0.6
0.8
1
Ix(c) TM TE
kx (k0)
In ensi y (a b. uni s)
qDi ac ion O de s a Fou ie Plane
4
5
6
7
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
-0.5-0.4-0.3-0.2 0.50.40.3
0.20.1-0.1 0
Fig. 4. MEMS-BMS BDG wi h econ igu able di ac ion o de s: expe imen . (a) Scanning elec on mic oscopy (SEM) images o MS1 and BMS on he
glass subs a e. (b) Measu ed di ac ion in ensi ies as a unc ion o ai gap sizes Taa he Fou ie plane o he inciden wa eleng h o λ=850 nm unde
TM exci a ion. (c) Ex ac ed di ac ion in ensi ies o di e en Fab y–Pe o o de s o q=q1=q2=4, 5, 6, 7 a wo speci ic ai gaps ha ac i ely chan-
nel he e lec ed powe be ween he +1 and −1 di ac ion o de s. (d) Measu ed di ac ion e iciencies o he −1 (black) and +1 ( ed) o de s a he wa e-
leng h o λ=850 nm as a unc ion o Ta o bo h TM and TE exci a ions. Di ac ion con as be ween ±1 o de s is also shown (g een). (e) Response ime
o swi ching be ween ±1 di ac ion o de s (q=q1=q2=4), measu ed by ac ua ing he inne elec odes o he MEMS mi o wi h a pe iodic ec angle
signal be ween 9 and 12 V.
applying ol ages up o 23 V, was ound o be om 1.7 o 3.5 µm,
co e ing ou adjacen Fab y–Pe o o de s (q1=q2=4, 5, 6, 7) a
he design wa eleng h o λ=800 nm (Fig. S25).
To cha ac e ize he MEMS-BMS BDG, we buil an op ical
se up comp ising a ibe -coupled supe con inuum lase wi h
pola iza ion op ics (linea pola ize and hal -wa e pla e) and a
CMOS came a o cap u ing di ec /Fou ie images, o al e -
na i ely using a spec og aph o wa eleng h- esol ed Fou ie
plane imaging (see Appendix A and Fig. S26). The MEMS mi o
was elec ically ac ua ed o modula e he op ical esponse o he
MEMS-BMS BDG, obse ed isually a he Fou ie plane wi h
di e en di ac ion o de s well sepa a ed (|m| ≤ 2), as shown
in Figs. 4(b) and 4(c). Fo bo h pola iza ions, i is clea ly seen
ha he blazed di ec ion can be swi ched be ween +1 and −1
di ac ion o de s in a pe iodic manne [Figs. 4(b) and 4(c)]. The
expe imen al di ac ion e iciencies o −1/+1 o de s a 850 nm
wa eleng h a e ∼30%/25% (q1=q2=4), and he con as
be ween ±1 o de s can each ∼ + 0.80/−0.75, indica ing p o-
nounced swi ching be ween ±1 di ac ion o de s [Fig. 4(d)]. The
high-con as swi chable di ac ion o de s, induced by ac ua ing
he inne elec odes o he MEMS mi o wi h 1-Hz al e na ing
9/11 V ol ages, a e also documen ed in a ideo cap u ed by he
CMOS came a (Visualiza ion 3). No e ha he op imum ope a-
ion wa eleng h shi ed om he designed wa eleng h o 800 nm
o ha o 850 nm wi h e iciencies no ably lowe han simula ion
p edic ions [Fig. 3( )] due o inc eased loss and su ace oughness
o he polyc ys alline gold me a-a oms [20,31] and he de ia ions
in SU-8 hickness and me a-a om sizes in ab ica ion (Figs. S15 o
S17, S27). I should also be no ed ha , s a ing om he maximum
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1562
o +1 di ac ion o de , he majo adia ion powe swi ches suc-
cessi ely o 0 and −1 di ac ion o de s when inc easing he ai gap
[Fig. 4(b)], scanning in s eps o e h ee dis inc di ac ion angles
(Visualiza ion 4). Combining his e ec wi h he MS design o
quasi-con inuous beam s ee ing be ween h ee (ze o h, i s , and
second) di ac ion o de s [20] may enable beam s ee ing o e i e
dis inc di ac ion angles co e ing p ac ically any angula ange
wi hin ±90◦, a unc ionali y ha migh be use ul o , e.g., LIDAR
applica ions.
The MEMS-BMS BDG pe o mance a o he inciden
wa eleng hs was also s udied, showing he beha io expec ed
om simula ions (see he p eceding sec ion). Fo example, he
+1 di ac ion o de exhibi s be e obus ness o wa eleng h
changes (Fig. S28), eaching e iciencies/con as s o ∼25%/+0.8
and 22%/+0.8 o inciden wa eleng hs o 800 and 900 nm,
espec i ely, whe eas he −1 o de exhibi s s onge wa eleng h
dependence, wi h es ima ed e iciencies/con as o ∼18%/−0.4
and 28%/–0.7. Addi ionally, he swi ching speed was also cha ac-
e ized by ac ua ing he MEMS mi o wi h a pe iodic ec angle
signal and de ec ing spa ially sepa a ed ±1 o de s, showing
ema kably as ise/ all imes o ∼5µs [Fig. 4(e)], which is ∼2
o de s o magni ude as e han ou p e ious MEMS- unable
single-laye MSs [20,21,23,24].
E. MEMS-BMS Vo ex Phase Pla e
The MEMS-BMS VPP o ealizing dynamically con olled
pola iza ion-independen 2D o ex beam gene a ion wi h
econ igu able opological cha ges o l= +1 and −1 [Fig. 5(a)]
is designed by selec ing ou MEMS-BMS cells (Fig. 2and
Supplemen 1 S15), wi h desi ed e lec ion phase and high ampli-
udes| | ≥ 0.8a wo di e en ai gaps [Figs.5(b) and 5(c)]. He e,
we ema k ha bo h MEMS-BMS VPP and BDG we e ab ica ed
on he same glass subs a e using iden ical p ocesses and assembled
in o a single MEMS-BMS componen . SEM images we e aken
a e MS1 and BMS ab ica ions [Fig. 5(d)], showing gene ally
good quali y, wi h me a-a om geome ies mi o ing he design
and nea -pe ec alignmen be ween he wo MS laye s, despi e he
p esence o ounded co ne s in he c oss-shaped me a-a oms.
To cha ac e ize he MEMS-BMS VPP, we added a e e ence
a m o ou expe imen al se up, allowing di ec in es iga ion o he
e lec ed phase p o iles (Fig. S31). The e e ence Gaussian beam
can be adjus ed o pe o m bo h co-axis and o -axis in e e om-
e y wi h he beam e lec ed om MEMS-BMS VPP. Fo bo h
pola iza ions, he gene a ion o a o ex beam wi h swi chable
opological cha ges o l= ±1 is clea ly obse ed in he in e e -
ence pa e ns. Wi h he MEMS mi o ac ua ed a wo di e en
ol ages o 14 and 17 V, nea - ield on-axis in e e og ams dis-
play spi al-shaped inges wi h opposi e winding, while a - ield
o -axis in e e og ams e eal o k-shaped inges wi h opposi e
o ien a ions [Fig. 5(e), Visualiza ion 5 and Visualiza ion 6]. The
o ex beam cha ac e ized by a opological cha ge o l= −1 wi h
Ta1=∼ q1×λ/2, exhibi s be e quali y [Figs. 5(e) and S30] as
only MS1 laye con ibu es o he o e all MEMS-BMS esponse
in his s a e. Due o he non-uni o m e lec ion ampli udes o
he MEMS-BMS uni cells [Fig. 5(c)] and ou quad an phase
disc e iza ion, he in ensi y pa e ns display asymme ic p o iles,
which a e also obse ed in simula ions (Fig. S30). Gi en he pos-
sibili ies o designing ine phase s eps using low-loss dielec ic
me a-a oms [30], he e iciency o he MEMS-BMS VPP and he
quali y o he esul ing o ex beam could be imp o ed.
3. DISCUSSION AND CONCLUSION
We ha e de eloped he elec ically d i en dynamic MEMS-BMS
pla o m wi h as esponse (∼5µs) by combining ligh weigh
piezoelec ic MEMS mi o s (see Appendix A) wi h plasmonic
bilaye MSs. This pla o m accommoda es unable opological
singula i ies in a 3D pa ame e space de ined by he esonance
p ope ies o he me a-a oms in bo h MS laye s and he MEMS-
BMS sepa a ion, enabling he eby comple e e lec ion phase
ans o ma ion and swi ching be ween wo encoded unc ion-
ali ies by ac ua ing he MEMS mi o . We ha e designed and
expe imen ally demons a ed MEMS-BMS componen s ope -
a ing in he nea -in a ed wa eleng h egime (∼800 nm) o
econ igu able BDG and VPP, bo h showing dis inc op ical
esponses a wo ope a ing s a es, speci ically wo MEMS-
BMS sepa a ions. The expe imen al di ac ion e iciencies o
∼30%/25% we e ealized wi h he MEMS-BMS BDG o espec-
i e −1/+1 di ac ion o de s along wi h a high con as o
∼ + 0.80/−0.75, con i ming he opposi e 1D phase g adien s
achie ed as expec ed om he simula ions. No e ha , acco d-
ing o ou simula ions, he e iciencies o >60% and con as o
>±0.9 can be ealized gi en mo e p ecise nano ab ica ion. Fo
he MEMS-BMS VPP, he gene a ion o he swi chable o ex
beam wi h econ igu able opological cha ges o l= ±1 was e i-
denced by acking he e lec ion in ensi y and 2D phase p o iles.
Since he dual-s a e ope a ion is achie ed using only wo se s o
ol ages, he MEMS-BMS can be scaled up o la ge ape u e sizes
while main aining small pixel sizes o op imal op ical esponses,
wi hou inc easing he con ol complexi y. To u he enhance
he MEMS-BMS pe o mance, expanding he me a-a om lib a y
wi h di e se ma e ials, geome ies, shapes, and o ien a ions in
each MS laye could p o ide mo e comple e ampli ude/phase
ans o ma ions and addi ional mul iplexing channels, le e aging
he design’s inhe en lexibili y. Ano he in iguing di ec ion
would be o u he exploi he unable MEMS-MS pla o m o
dedica ed in es iga ions o eme gence, e olu ion, and annihila ion
o he opological phase singula i ies [24,60–63]. The subs an-
ial design eedom and ine- uning capabili ies o he de eloped
pla o m enable s udies o opological singula i ies in a high-
dimensional pa ame e space, pa ing he way o mul i unc ional
and mul iplexing unable opological me a-op ics.
APPENDIX A: METHODS
1. Nume ical Calcula ions
The calcula ion o he complex e lec ion coe icien o he
MEMS-BMS uni cell [Fig. 2(a)] is conduc ed in wo dis inc
s eps: i s , a simula ion model is cons uc ed in COMSOL
Mul iphysics 5.6. This model consis s o a gold me a-a om placed
wi hin a squa e uni cell wi h a side leng h o 3=300 nm. The
uppe /lowe hal es o he in ini e space a e occupied by glass/
SU-8 o he MS1 uni cell, o SU-8/ai ma e ials o he MS2
uni cell. Pe iodic bounda y condi ions we e applied in bo h x
and ydi ec ions; he eby he simula ion is ela ed o a 2D in ini e
a ay composed o pe iodically a anged MS uni cells. Wi h his
model, we calcula ed he complex e lec ion and ansmission
coe icien s o single-laye MS1 and MS2 uni cells wi h a ying
me a-a om geome ies and dimensions, o no mally inciden
ligh om ei he he uppe o lowe side. No e ha each co ne
o he c oss-shaped me a-a oms is ounded wi h a adius o 5 nm
in he COMSOL simula ions. Subsequen ly, we in eg a ed he
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1563
Inciden
Beam
l=−1l=+1
(a)
VMEMS=Va2
VMEMS=Va1
MEMS-BMS VPP
(c)(b)
ϕ(x,y,Ta2)
0
0.5ππ
1.5π
ϕ(x,y,Ta1)
0
0.5π
π
1.5π
BMS
MEMS
x(TM)
z
y(TE)
k
Va1 Va2
x
y
x
y
(d)
0.5 µm
Λ
Λ
MS1
MS2
1
∠
-200°
-100°
0°
100°
200°
| |
Uni cells o MEMS-BMS VPP
0.9
0.8
0.7
0.6
0.5
Λ
(e)
TM
17 V (l=+1)
In e e og am(DI)In e e og am (FI)
TM TE TM TE
14 V (l=−1) Subs a e (l=0)
Vm
MEMS-BMS VPP
Pol.
I1
0.8
0.6
0.4
0.2
0
In ensi y (FI)
23 4
∠
| |
150+q2λ/2
Ta(nm)
350+q1λ/2
Glass
MS1
MS1
Glass
SU8
MS1
MS2
BMS
I1
0.8
0.6
0.4
0.2
0
I1
0.8
0.6
0.4
0.2
0
Fig. 5. MEMS-BMS VPP o gene a ing o ex beam wi h swi chable opological cha ges. (a) Schema ic ende ing o he MEMS-BMS VPP o gene -
a ing o ex beams wi h swi chable opological cha ges o l= −1 and +1 when he MEMS-BMS is elec ically econ igu ed be ween wo ai gaps wi h di -
e en ol ages. (b) Tunable spi al phase dis ibu ions o swi chable o ex beams wi h l= −1 and +1 a espec i e ai gaps o Ta=(350 +q1×λ/2)and
(150 +q2×λ/2)nm. (c) Calcula ed e lec ion ampli udes (black ma ke s) and phases ( ed ma ke s) o ou selec ed uni cells o cons uc ing he MEMS-
BMS VPP. (d) SEM images o he MS1 and BMS on he glass subs a e. (e) Measu ed on-axis in e e og ams a he di ec image plane (scale ba , 20 µm),
o -axis in e e og ams a he Fou ie plane (scale ba , 0.1k0,k0=2π/λ), and he in ensi y p o iles a he Fou ie plane (scale ba , 0.1k0,k0=2π/λ) o he
gene a ed o ex beams a he wa eleng hs o λ=750 nm o bo h TM and TE exci a ions. The o ex beam’s opological cha ge swi ches om l= −1 o
+1 by ac ua ing he inne elec odes o he MEMS mi o om 14 o 17 V, espec i ely. The in e e og ams a e p oduced by supe posing he o ex beam
wi h a e e ence Gaussian beam.
MS1/MS2 uni cells wi h he MEMS gold mi o using he ans-
e ma ix me hod and calcula ed he o al complex e lec ion
coe icien s o he MEMS-BMS uni cell (Supplemen 1 S2 and S3,
Figs. S2 o S6). To a oid any unwan ed nea - ield coupling e ec s
be ween MS laye s, he hickness o he SU-8 laye be ween MS1
and MS2 laye s is inc eased om 136 nm (i.e., OLBMS =λe /4)
o 389 nm (i.e., OLBMS =3λe /4), o a design wa eleng h o
800 nm (Supplemen 1 S5 and S6).
Wi h he abo e-de eloped app oach, we calcula ed he phase
esponse maps o Ta=150 +q2×λ/2 and 350 +q1×λ/2
(q1and q2a e non-nega i e in ege s) wi h MS1/MS2 uni cells
comp ising a ious me a-a om geome ies and sizes (Fig. 2and
Figs. S5, S6). To design he MEMS-BMS BDG wi h econ-
igu able di ac ion o de s, we selec ed 12 MEMS-BMS uni
cells, each ea u ing a e lec ion ampli ude ≥0.7, om he phase
maps, o app oxima ing wo con as ing linea phase p o iles
(i.e., ±2π/123) a wo dis inc ai gaps. The en i e MEMS-
BMS BDG a angemen consis s o pe iodically a anged BDG
supe cells along bo h he xand ydi ec ions, wi h pe iods o
123=3.6 µm along he xdi ec ion and 3=0.3 µm along
he ydi ec ion. In simula ion, we applied pe iodic bounda y
condi ions along bo h xand ydi ec ions, co esponding o a
pe iodically a anged 2D in ini e a ay. A no mally inciden
plane wa e was applied, and he e lec ed ligh di ec ed o di -
e en di ac ion o de s was moni o ed, wi h di e en ai gaps
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1564
o es ima ing he dynamic di ac ion e iciencies and con-
as . Using his model, he MEMS-BMS BDG ope a ion was
in es iga ed unde di e en inciden wa eleng hs [Fig. 3( )
and Fig. S14], inc eased loss o he gold me a-a oms (Fig. S15),
a ia ions in SU-8 hickness (Fig. S16), de ia ions in me a-
a om sizes (Fig S17), and he ole ance o he angle o incidence
(Fig. S18).
The MEMS-BMS VPP o o ex beam gene a ion wi h
swi chable opological cha ges was designed in a simila ashion. In
his design, we selec ed ou MEMS-BMS uni cells, each ea u ing
a high- e lec ion ampli ude ≥0.8, o achie e a unable 2D spi al
phase swi ching be ween 2πand −2πa wo di e en ai gaps.
The en i e MEMS-BMS VPP s uc u e was simula ed wi h he 3D
ini e-di e ence ime-domain (FDTD) me hod. Due o he high
compu a ional demands, we scaled down he o e all size o he
MEMS-VPP o a diame e o 3µm and employeda Gaussianbeam
wi h a 1.2-µm wais adius as he inciden ligh in simula ion.
Re lec ed ields we e moni o ed and subsequen ly p ojec ed o
he a ield, o isualizing he in ensi y and phase p o ile a he
Fou ie plane (Fig. S30).
2. Fab ica ion
The BMSs o MEMS-BMS componen s we e ab ica ed by
epea ing he s anda d elec on-beam li hog aphy (EBL),
gold deposi ion, and li -o p ocesses wice. Addi ionally, an
SU-8 space laye was added a e he ab ica ion o he MS1
laye (Supplemen 1 S12). Fo he ab ica ion o each MS
laye , he p ocess includes se e al s eps: i s , a 100-nm- hick
poly(me hyl me hac yla e) (PMMA A2, Mic oChem) laye
and a 40-nm- hick conduc i e polyme laye (AR-PC 5090,
All esis ) we e successi ely spin-coa ed on a squa e glass sub-
s a e (14 mm ×14 mm ×400 µm). Then, he MS pa e n was
de ined a he cen al a ea o he glass subs a e using EBL (JEOL
JSM-6500F ield-emission SEM wi h a Rai h Elphy Quan um
li hog aphy sys em). A e de elopmen , he MS laye was o mu-
la ed by deposi ing a 1-nm Ti adhesion laye and a 50-nm gold
laye (To nado 400, C yo ox) ollowed by li -o in ace one. No e
ha he alignmen ma ke s we e also ab ica ed concu en ly wi h
he MS1 laye . This is c ucial as i acili a es he alignmen p ocess
du ing he ab ica ion o he MS2 laye . A e he ab ica ion o he
MS1 laye , an SU-8 laye , aiming o a a ge hickness o 389 nm,
was spin-coa ed o e he en i e glass subs a e, ollowed by a so
bake, UV cu ing, and ha d bake p ocess o make i a pe manen
space laye . Finally, he MS2 laye was ab ica ed ollowing he
same s eps as hose used o he MS1 laye , wi h an addi ional
alignmen p ocess included.
Compa ed wi h ou ea lie wo k, his MEMS mi o u ilizes
he same ab ica ion p ocess [20,21], bu wi h a di e en elec ode
layou and smalle mi o diame e [Fig. 1(d) and Fig. S19]. The
smalle mi o diame e o 500 µm educes mi o mass, esul ing
in highe esonance equency and as e ac ua ion. The piezo-
elec ic memb ane esponsible o mo ing he mi o consis s o
ou can ile e s, ins ead o he annula memb ane used ea lie .
The s i ness o hese can ile e s is less han ha o he annula
memb ane, bu oge he wi h he smalle mass o he mi o , he
o al esponse is as e while s ill ha ing a displacemen ange o
∼1µm o 23 V. This is adequa e o ope a ing ou MEMS-BMS
componen be ween wo con igu a ion s a es, which equi es
∼200 nm, as well as o in es iga ing he MEMS-BMS beha io s
ac oss se e al adjacen Fab y–Pe o o de s ha exhibi a pe iod o
λ/2=400 nm o inciden ligh wi h λ=800 nm. Be o e assem-
bling he MEMS-BMS componen , he ul a- la piezoelec ic
MEMS mi o was spu e ed wi h a 100-nm- hick gold laye o
block any ansmission a ound he nea -in a ed wa eleng h o
λ=800 nm.
3. Cha ac e iza ion
The expe imen al se up o MEMS-BMS BDG is shown in Fig.
S26. A collima ed ibe -coupled supe con inuum lase (Supe K
Ex eme, NKT) was di ec ed h ough a i s hal -wa e pla e
(HWP1, AHWP10M-980, Tho labs), an ND il e , a mi o ,
a linea pola ize (LP; LPNIR050-MP2, Tho labs), a second
hal -wa e pla e (HWP2, AHWP10M-980, Tho labs), a lens
L1( =200 mm), wo beam spli e s (BS1,2; CCM1-BS014,
Tho labs) successi ely, and hen ocused on o he MEMS-BMS
BDG by an objec i e (Obj; M Plan Apo, ×50/0.55, Mi u oyo).
The combina ion o LP and HWP2is used o adjus ing he inpu
linea pola iza ion o ien a ion while keeping a cons an inciden
powe . The e lec ed ligh was collec ed by he same objec i e and
passed h ough BS2and a ube lens (TL; TTL200-S8, Tho labs),
gene a ing he i s di ec image whe e an i is is placed o il e ing
ou he e lec ed ligh wi hin he in e es ed a ea in he MEMS-
BMS componen . The i s di ec image is hen ans o med by
a elay lens (RL; AC254-200-B-ML, =200 mm, Tho labs)
o he co esponding Fou ie image and cap u ed by a CMOS
came a (CMOS; DCC1545M, Tho labs), acco ding o a 2
con igu a ion. A lip lens (FL; AC254-100-B-ML, =100 mm,
Tho labs) is used o swi ching be ween he di ec and Fou ie
images. Addi ionally, a lip mi o (M2) is used o swi ch be ween
he CMOS came a and spec og aph (Model SR-303i-A-SIL,
Ando ), o di ec /Fou ie plane imaging o wa eleng h- esol ed
Fou ie plane imaging, espec i ely.
To cha ac e ize he MEMS-BMS componen s, a no mally
inciden Gaussian beam was sligh ly ocused in o a spo wi h
a diame e o ∼30 µm, which is compa able o he ab ica ed
MEMS-BMS sizes o 28.8 µm×28.8 µm o he BDG and
30 µm×30 µm o he VPP. This beam spo size co esponds
o a beam di e gence o ∼0.7◦, es ima ed using θ=λ/(πnw0),
whe e θis he beam di e gence, λis he ee-space wa eleng h
(800 nm), nis he e ac i e index o he medium (1.466), and
w0is he beam adius (15 µm). No e ha an i is is used in he
in e media e image plane o il e ou he a ea o in e es in
he measu emen , as shown in Fig. S26C and Fig. S31B. The
expe imen ally ob ained di ac ion con as is calcula ed by
[I(m= +1)−I(m= −1)]/[I(m= +1)+I(m= −1)].
To es ima e he swi ching speed o he MEMS-BMS BDG
be ween ±1 di ac ion o de s, he se up desc ibed abo e is modi-
ied by eplacing he supe con inuum lase and CMOS came a
wi h a CW Ti: sapphi e lase (Spec a-Physics 3900S, wa eleng h
ange: 700–1000 nm), and a pho ode ec o (PD; PDA20CS-EC,
Tho labs), espec i ely. The signals om he PD a e acqui ed
wi h an oscilloscope (DSOX2024A, Keysigh ). In he mea-
su emen , he MEMS-BMS BDG is ac ua ed wi h pe iodically
al e na ing ol age signals om a unc ion gene a o (TOE 7402,
TOELLNER).
Fo cha ac e izing MEMS-BMS VPP, we added a e e ence
a m o he abo e-desc ibed se up (Fig. S31). The modi ied se up
esembles a Michelson in e e ome e in which he e e ence
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1565
Gaussian beam can be adjus ed o pe o m on-axis and o -axis
in e e ome y wi h he beam e lec ed om he MEMS-BMS
VPP. The e e ence a m in oduced a e e ence Gaussian beam
wi h a simila in ensi y and nea ly equal op ical pa h leng h o ha
o he e lec ed ligh om he MEMS-BMS. The in e e og am
p o iles a he di ec image and Fou ie planes can be cap u ed by a
CMOS came a, o es ima ing he opological cha ges o he gen-
e a ed o ex beam. The in ensi y p o iles a he di ec and Fou ie
image planes can be eco ded wi h he e e ence beam blocked.
Funding. ATTRACT p og amme unded by he Eu opean Union’s Ho izon
2020 Resea ch and Inno a ion P og amme (101004462); Villum Fonden
(37372, 50343, awa d in Technical and Na u al Sciences 2019); Danma ks F ie
Fo sknings ond (1134-00010B); No ges Fo sknings åd (323322).
Acknowledgmen . We g a e ully acknowledge Ka olina Milenko and
Zeljko Skokic a SINTEF o designing and ab ica ing he MEMS mi o s,
Shailesh Kuma o help se ing up he spec og aph, To gom Yezekyan and
Volodymy Zenin o assis ance wi h expe imen se up, as well as Zhengli Han
o ui ul discussions. C.M. and S.I.B. concei ed he idea. C.M. pe o med he
simula ions, ab ica ed he BMS samples, buil he se up, conduc ed he measu e-
men s, and analyzed he da a. P.C.V.T. assembled MEMS-BMS de ices. C.M.
and F.D. p o ided he i s d a o he manusc ip . All au ho s con ibu ed o he
discussion o he esul s ob ained and w i ing he manusc ip . S.I.B. supe ised he
p ojec .
Disclosu es. The au ho s decla e no con lic s o in e es .
Da a a ailabili y. All da a a e a ailable in he manusc ip o he supplemen-
a y ma e ials.
Supplemen al documen . See Supplemen 1 o suppo ing con en .
REFERENCES
1. J. Yang, S. Gu ung, S. Bej, e al., “Ac i e op ical me asu aces: comp e-
hensi e e iew on physics, mechanisms, and p ospec i e applica ions,”
Rep. P og. Phys. 85, 036101 (2022).
2. O. A. M. Abdel aou , Z. Wang, H. Liu, e al., “Recen ad ances in un-
able me asu aces: ma e ials, design, and applica ions,” ACS Nano 16,
13339–13369 (2022).
3. T. Gu, H. J. Kim, C. Ri e o-Baleine, e al., “Recon igu able me asu aces
owa ds comme cial success,” Na . Pho onics 17, 48–58 (2023).
4. A. M. Shal ou , V. M. Shalae , and M. L. B onge sma, “Spa io empo al
ligh con ol wi h ac i e me asu aces,” Science 364, eaa 3100 (2019).
5. Q. He, S. Sun, and L. Zhou, “Tunable/ econ igu able me asu aces:
physics and applica ions,” Resea ch 2019, 1849272 (2019).
6. Y.-C. Ling and S. J. B. Yoo, “Re iew: unable nanopho onic me as uc-
u es,” Nanopho onics 12, 3851–3870 (2023).
7. S.-Q. Li, X. Xu, R. M. Vee il, e al., “Phase-only ansmissi e spa ial ligh
modula o based on unable dielec ic me asu ace,” Science 364,
1087–1090 (2019).
8. P. C. Wu, R. A. Pala, G. K. Shi manesh, e al., “Dynamic beam s ee -
ing wi h all-dielec ic elec o-op ic III–V mul iple-quan um-well
me asu aces,” Na . Commun. 10, 3654 (2019).
9. G. K. Shi manesh, R. Sokhoyan, P. C. Wu, e al., “Elec o-op ically un-
able mul i unc ional me asu aces,” ACS Nano 14, 6912–6920 (2020).
10. J. Pa k, B. G. Jeong, S. I. Kim, e al., “All-solid-s a e spa ial ligh modula-
o wi h independen phase and ampli ude con ol o h ee-dimensional
LiDAR applica ions,” Na . Nano echnol. 16, 69–76 (2021).
11. P. Moi a, X. Xu, R. M. Vee il, e al., “Elec ically unable e lec i e me a-
su aces wi h con inuous and ull-phase modula ion o high-e iciency
wa e on con ol a isible equencies,” ACS Nano 17, 16952–16959
(2023).
12. Z. Fang, R. Chen, J. E. F öch, e al., “Non ola ile phase-only ansmis-
si e spa ial ligh modula o wi h elec ical add essabili y o indi idual
pixels,” ACS Nano 18, 11245–11256 (2024).
13. J. Sisle , P. Thu eja, M. Y. G ajowe , e al., “Elec ically unable space–
ime me asu aces a op ical equencies,” Na . Nano echnol. 19,
1491–1498 (2024).
14. M. Decke , I. S aude, M. Falkne , e al., “High-e iciency dielec ic
Huygens’ su aces,” Ad . Op . Ma e . 3, 813–820 (2015).
15. Q. Song, M. Odeh, J. Zúñiga-Pé ez, e al., “Plasmonic opological me a-
su ace by enci cling an excep ional poin ,” Science 373, 1133–1137
(2021).
16. X. Wu, J. Zhu, F. Lin, e al., “S udy o a high-index dielec ic non-
He mi ian me asu ace and i s applica ion in holog ams,” ACS Omega 7,
44743–44749 (2022).
17. P. Fu, S. Du, W. Lan, e al., “Deep lea ning enabled opological design o
excep ional poin s o mul i-op ical-pa ame e con ol,” Commun. Phys.
6, 254 (2023).
18. X. Wu, Z. Feng, X. Zhao, e al., “Full-space and mul i-pola iza ion holo-
g ams ealized by a non-He mi ian bilaye me asu ace,” Op . Le . 48,
6044–6047 (2023).
19. Z. Yang, P. Huang, Y. Lin, e al., “C ea ing pai s o excep ional poin s
o a bi a y pola iza ion con ol: asymme ic ec o ial wa e on
modula ion,” Na . Commun. 15, 232 (2024).
20. C. Meng, P. C. V. Th ane, F. Ding, e al., “Dynamic piezoelec ic MEMS-
based op ical me asu aces,” Sci. Ad . 7, eabg5639 (2021).
21. C. Meng, P. C. V. Th ane, F. Ding, e al., “Full- ange bi e ingence con ol
wi h piezoelec ic MEMS-based me asu aces,” Na . Commun. 13, 2071
(2022).
22. C. A. Di dal, P. C. V. Th ane, F. T. Dullo, e al., “MEMS- unable dielec ic
me asu ace lens using hin- ilm PZT o la ge displacemen s a low ol -
ages,” Op . Le . 47, 1049–1052 (2022).
23. Y. Deng, C. Meng, P. C. V. Th ane, e al., “MEMS-in eg a ed me a-
su aces o dynamic linea pola ize s,” Op ica 11, 326–332
(2024).
24. F. Ding, Y. Deng, C. Meng, e al., “Elec ically unable opological phase
ansi ion in non-He mi ian op ical MEMS me asu aces,” Sci. Ad . 10,
eadl4661 (2024).
25. E. A babi, A. A babi, S. M. Kamali, e al., “MEMS- unable dielec ic
me asu ace lens,” Na . Commun. 9, 812 (2018).
26. Z. Han, S. Colbu n, A. Majumda , e al., “MEMS-ac ua ed me asu ace
Al a ez lens,” Mic osys . Nanoeng. 6, 79 (2020).
27. Z. Han, S. Colbu n, A. Majumda , e al., “Millime e -scale ocal leng h
uning wi h MEMS-in eg a ed me a-op ics employing high- h oughpu
ab ica ion,” Sci. Rep. 12, 5385 (2022).
28. Z. Han, C. F ydendahl, N. Mazu ski, e al., “MEMS can ile e –con olled
plasmonic colo s o sus ainable op ical displays,” Sci. Ad . 8, eabn0889
(2022).
29. M. Liu, M. Susli, D. Sil a, e al., “Ul a hin unable e ahe z abso be
based on MEMS-d i en me ama e ial,” Mic osys . Nanoeng. 3, 17033
(2017).
30. A. L. Hols een, A. F. Cihan, and M. L. B onge sma, “Tempo al colo mix-
ing and dynamic beam shaping wi h silicon me asu aces,” Science 365,
257–260 (2019).
31. P. C. V. Th ane, C. Meng, F. Ding, e al., “MEMS unable me asu aces
based on gap plasmon o Fab y-Pé o esonances,” Nano Le . 22,
6951–6957 (2022).
32. C. P ei e , C. Zhang, V. Ray, e al., “High pe o mance bianiso opic
me asu aces: asymme ic ansmission o ligh ,” Phys. Re . Le . 113,
023902 (2014).
33. C. Menzel, J. Spe hake, and T. Pe sch, “E icien ea men o s acked
me asu aces o op imizing and enhancing he ange o accessible op i-
cal unc ionali ies,” Phys. Re . A 93, 063832 (2016).
34. Y. Zhou, I. I. K a chenko, H. Wang, e al., “Mul i unc ional me aop ics
based on bilaye me asu aces,” Ligh Sci. Appl. 8, 80 (2019).
35. Z. Han, S. Ohno, and H. Minamide, “Spec al phase singula i y in a
ansmission- ype double-laye me ama e ial,” Op ica 7, 1721–1728
(2020).
36. X. Cai, R. Tang, H. Zhou, e al., “Dynamically con olling e ahe z wa e-
on s wi h cascaded me asu aces,” Ad . Pho on. 3, 036003 (2021).
37. Y. Wang, Q. Fan, and T. Xu, “Design o high e iciency ach oma ic me -
alens wi h la ge ope a ion bandwid h using bilaye a chi ec u e,” Op o-
Elec on. Ad . 4, 200008 (2021).
38. P. Geo gi, Q. Wei, B. Sain, e al., “Op ical sec e sha ing wi h cascaded
me asu ace holog aphy,” Sci. Ad . 7, eab 9718 (2021).
39. H. Zheng, M. He, Y. Zhou, e al., “Compound me a-op ics o comple e
and loss-less ield con ol,” ACS Nano 16, 15100–15107 (2022).
40. L. Deng, Z. Li, Z. Zhou, e al., “Bilaye -me asu ace design, ab ica ion,
and unc ionaliza ion o ull-space ligh manipula ion,” Ad . Op . Ma e .
10, 2102179 (2022).
41. S. Gao, C. Zhou, W. Liu, e al., “Dielec ic pola iza ion- il e ing me asu -
ace double o i unc ional con ol o ull-space isible ligh ,” Lase
Pho on. Re . 16, 2100603 (2022).
Resea ch A icle Vol. 11, No. 11 / No embe 2024 / Op ica 1566
42. A. Be khou and A. F. Koende ink, “A simple ans e -ma ix model o
me asu ace mul ilaye sys ems,” Nanopho onics 9, 3985–4007 (2020).
43. M. Liu, C. Zhao, Y. Zeng, e al., “E olu ion and non ecip oci y o loss-
induced opological phase singula i y pai s,” Phys. Re . Le . 127,
266101 (2021).
44. G. E molae , K. Vo onin, D. G. Ba ano , e al., “Topological phase singu-
la i ies in a omically hin high- e ac i e-index ma e ials,” Na . Commun.
13, 2049 (2022).
45. R. Colom, E. Mikhee a, K. Achou i, e al., “C ossing o he b anch
cu : he opological o igin o a uni e sal 2π-phase e a da ion in
non-He mi ian me asu aces,” Lase Pho on. Re . 17, 2200976 (2023).
46. E. Mikhee a, R. Colom, K. Achou i, e al., “Asymme ic phase modu-
la ion o ligh wi h pa i y-symme y b oken me asu aces,” Op ica 10,
1287–1294 (2023).
47. M. Liu, W. Chen, G. Hu, e al., “Spec al phase singula i y and opological
beha io in pe ec abso p ion,” Phys. Re . B 107, L241403 (2023).
48. A. Be khou and A. F. Koende ink, “Pe ec abso p ion and phase singu-
la i ies in plasmon an enna a ay e alons,” ACS Pho on. 6, 2917–2925
(2019).
49. Z. Sako ic, A. Wa e, M. Po inelli, e al., “Pe ec abso p ion a he
ul ima e hickness limi in plana ilms,” ACS Pho on. 10, 4244–4251
(2023).
50. E. Cuswo h, V. G. K a e s, and A. N. G igo enko, “Topological da kness
in op ical he e os uc u es: p edic ion and con i ma ion,” ACS Pho on.
10, 3715–3722 (2023).
51. A. N. G igo enko, P. I. Niki in, and A. V. Kabashin, “Phase jumps and
in e e ome ic su ace plasmon esonance imaging,” Appl. Phys. Le .
75, 3917–3919 (1999).
52. V. G. K a e s, F. Schedin, R. Jalil, e al., “Singula phase nano-op ics in
plasmonic me ama e ials o label- ee single-molecule de ec ion,” Na .
Ma e . 12, 304–309 (2013).
53. Y. Wu, L. Kang, H. Bao, e al., “Exploi ing opological p ope ies o Mie-
esonance-based hyb id me asu aces o ul a as swi ching o ligh
pola iza ion,” ACS Pho on. 7, 2362–2373 (2020).
54. J. H. Pa k, A. Ndao, W. Cai, e al., “Symme y-b eaking-induced plas-
monic excep ional poin s and nanoscale sensing,” Na . Phys. 16,
462–468 (2020).
55. G. Ma, W. Shen, D. S. Sanchez, e al., “Exci ons enabled opological
phase singula i y in a single a omic laye ,” ACS Nano 17, 17751–17760
(2023).
56. Z. Sako ic, A. K asnok, A. Alú, e al., “Topological sca e ing singula i ies
and embedded eigens a es o pola iza ion con ol and sensing applica-
ions,” Pho on. Res. 9, 1310–1323 (2021).
57. T. Liu, S. Ma, B. Yang, e al., “E ec i e-medium heo y o mul ilaye
me ama e ials: ole o nea - ield co ec ions,” Phys. Re . B 102, 174208
(2020).
58. B. Mi zapou beinekalaye, A. McClung, and A. A babi, “Gene al lossless
pola iza ion and phase ans o ma ion using bilaye me asu aces,” Ad .
Op . Ma e . 10, 2102591 (2022).
59. A. Po s and S. I. Bozhe olnyi, “Plasmonic me asu aces o e icien
phase con ol in e lec ion,” Op . Exp ess 21, 27438–27451 (2013).
60. Z. Sako ic, P. S anko ic, V. Bengin, e al., “Non-He mi ian con ol o
opological sca e ing singula i ies eme ging om bound s a es in he
con inuum,” Lase Pho on. Re . 17, 2200308 (2023).
61. C. Guo, J. Li, M. Xiao, e al., “Singula opology o sca e ing ma ices,”
Phys. Re . B 108, 155418 (2023).
62. A. K asnok, D. Ba ano , H. Li, e al., “Anomalies in ligh sca e ing,” Ad .
Op . Pho on. 11, 892–951 (2019).
63. F. Mon icone, H. M. Doeleman, W. Den Hollande , e al., “T apping ligh
in plain sigh : embedded pho onic eigens a es in ze o-index me ama e-
ials,” Lase Pho on. Re . 12, 1700220 (2018).
94 CHAPTER 5. ARTICLES
5.5 Me asu ace Pola ime e o S uc u al Imag-
ing and Tissue Diagnos ics
Th ane P.C.V., Meng C., Byko A., Sie yi O., Ding F., Meglinski I., Di dal
C.A. and Bozhe olnyi S.I.
Me asu ace Pola ime e o S uc u al Imaging and Tissue Diagnos ics.
Submi ed, p ep in a h ps://doi.o g/10.48550/a Xi .2501.05864 (2025).
Me asu ace Pola ime e o S uc u al Imaging and
Tissue Diagnos ics
Paul Th ane1,2*†, Chao Meng2*†, Alexande Byko 3, Oleksii Sie yi3, Fei Ding2, Igo Meglinski4,
Ch is ophe A. Di dal1, Se gey I. Bozhe olnyi2*
1Sma Senso s and Mic osys ems, SINTEF Digi al, Gaus adalleen 23C, 0373 Oslo, No way.
2SDU Cen e o Nano Op ics, Uni e si y o Sou he n Denma k, Campus ej 55, DK-5230
Odense, Denma k.
3OPEM, ITEE, Uni e si y o Oulu, 90014 Oulu, Finland.
4College o Enginee ing and Physical Sciences, As on Uni e si y, Bi mingham B4 7ET, U.K.
* P.T: paul. h ane@sin e .no; C.M: [email p o ec ed].dk; S.I.B: [email protected]
† These au ho s con ibu ed equally
Abs ac
His opa hology, he s udy and diagnosis o disease h ough analysis o issue samples, is an
indispensable pa o mode n medicine. Howe e , he p ac ice is ime consuming and labo
in ensi e, compelling e o s o imp o e he p ocess and de elop new app oaches. One
pe spec i e echnique in ol es mapping changes in he pola iza ion s a e o ligh sca e ed by
he issue, bu he con en ional implemen a ion equi es bulky pola iza ion op ics and is slow.
We epo he design, ab ica ion and cha ac e iza ion o a compac me asu ace pola ime e
ope a ing a 640 nm enabling simul aneous de e mina ion o S okes pa ame e s and deg ee o
pola iza ion wi h ±2% accu acy. To alida e i s use o his opa hology we map pola iza ion s a e
changes in a issue phan om mimicking a biopsy wi h a cance ous inclusion, compa ing i o a
comme cial pola ime e . The esul s indica e a g ea po en ial and sugges se e al imp o emen s
wi h which we belie e me asu ace pola ime e based de ices will be eady o p ac ical
his opa hology applica ion in clinical en i onmen .
In oduc ion
Pola ime ic mapping and cha ac e iza ion o biological issues, including Muelle ma ix and
S okes imaging, ha e shown significan po en ial o applica ions in his ological issue
cha ac e iza ion1–3. Muelle ma ix imaging4, o example, has been ound e ec i e o he
de ec ion o mo phological and s uc u al al e a ions associa ed wi h cance 5,6, which has led o
he de elopmen s o sys ems such as Muelle ma ix imaging endoscopes7,8. Simila ly,
pola iza ion-holog aphic Muelle ma ix me hods9,10 can be applied o he assessmen o he 3D
mo phology o biological issues wi h applica ions in disease diagnosis11, while S okes
pola ime y has been demons a ed as an e ec i e ool o sc eening he p og ession o
Alzheime 's disease12. These label- ee and non-des uc i e imaging echniques can acili a e he
analysis o biological issue by elimina ing he need o con en ional sec ioning and s aining
p ocedu es. One such p omising echnique in ol es s uc u al imaging o issue samples by
mapping how he s a e o pola iza ion (SOP) o ligh is al e ed by sca e ing p ocesses in he
issue, and has been applied o he diagnosis and g ading o colon cance 13. P obing he biological
s uc u es in his way no only emo es he need o s ain he samples and can be done wi hou
sec ioning and moun ing bu is also sui able o au oma ed measu emen acquisi ion and
analysis. Such a de elopmen would significan ly speed up diagnosis and subsequen ea men ,
in addi ion o educing in a- and in e obse e a iabili y o diagnosis a ising om he s ong
eliance on pe sonal skills and qualifica ions o con en ional me hods.
C ucial o such pola iza ion-based measu emen s a e pola ime e s – de ices ha measu e he
ligh SOP. Pola ime e s all in o one o wo ca ego ies depending on hei p inciple o ope a ion.
Di ision-o - ime sys ems ha e a ime a ying componen ha measu es di e en p ope ies o
he incoming ligh a di e en imes, de e mining he ull SOP a e se e al measu emen s. One
common way o doing his is by sending he ligh h ough a o a ing qua e wa epla e wi h a s a ic
linea pola ize and hen measu ing he ansmi ed in ensi y, he Fou ie decomposi ion o his
signal is hen used o find he SOP14. Di ision-o -space sys ems, on he o he hand, spli he
incoming ligh in o se e al pa s wi h a ying in ensi ies dependen on he SOP, which can hen
be measu ed simul aneously and used o calcula e he ini ial SOP15,16. Bo h ca ego ies o
pola ime e s ha e seen majo de elopmen s in la e yea s, building on ad ances in in eg a ed
op ics o enable compac sys ems ha a e as e , mo e accu a e and enable new use cases17.
Nanos uc u ed su aces o simul aneous pola ime y ha e, o example, been used o make
Figu e 4. Cha ac e iza ion o he MS pola ime e wi h ully pola ized ligh . a Pho o o he calib a ion
se up, whe e a fibe coupled lase (λ = 640 nm) sends ligh h ough pola iza ion op ics and in o he MS
pola ime e . The in eg a ed CMOS came a has been emo ed in he pho o (see Figu e 1 d, which makes i
possible o check he beam alignmen on he MS wi h a sepa a e came a placed ou side he pola ime e
( op- igh o he pic u e). b and c Beam alignmen images. The di e ence be ween he wo images is he
wid h o he inciden beam. d Di ac ion pa e n acquisi ion. A e alignmen he CMOS came a is inse ed
in o he se up and acqui es di ac ion pa e n images. e Acquisi ion p ocess summa y: The a e age noise
alue is sub ac ed om he image, he di ac ion spo in ensi ies a e summed up and hese a e used o
calcula e he S okes ec o s by applying a calib a ion ma ix. Measu emen esul s o di e en
pola iza ion s a es wi h DOP = 1. Ci cles wi h he same colo and connec ed by a line co espond o he
same measu emen o a S okes ec o and co esponding DOP. The op ow p esen s measu emen s done
wi h a Tho labs PAX1000, he MS pola ime e be o e and a e calib a ion. The bo om ow displays he
a ia ion be ween measu emen s. g Subse o he pola iza ion s a e measu emen s plo ed on a Poinca e
sphe e. Ci cles a e measu ed by he Tho labs PAX1000, + signs a e om he MS pola ime e be o e
calib a ion and iangles a e a e calib a ion. Poin s co esponding o he same measu ed pola iza ion
s a e a e shown in he same colo , o mos measu emen s he ci cles and iangles o e lap – indica ing
good ag eemen a e calib a ion.
When he DOP is less han 1, he e is some ligh in ensi y in all he di ac ion spo s, e en o he
pola iza ion s a es ha he g a ings ha e been designed o , see example images in Figu e 5 a-c.
The plo s in Figu e 5 d-e a e DOP measu emen s o pa ially pola ized s a es close o he line
be ween s2 = 1 and s2 = -1, whe e he pola ime e is expec ed o ha e he wo s pe o mance due
o he une en e iciencies o he SCab g a ing. As can be obse ed in Figu e 5 e, o la ge DOP > 0.8
he di e ence be ween he e e ence pola ime e and he MS pola ime e a e calib a ion s ays
below 2%, while o lowe DOP(i.e., 0.8>DOP>0.2) he di e ence is below 5%. When he DOP
d ops below 0.2 he di e ence becomes much la ge , as he in ensi y a ia ions in he di ac ion
spo s become smalle . Fo his eason, eplacing he CMOS senso wi h a coun e pa ha has
be e in ensi y esolu ion and less noise should imp o e he accu acy o low DOP
measu emen s.
Figu e 5. Cha ac e iza ion o he MS pola ime e wi h pa ially pola ized ligh . a-c Th ee example
images o he di ac ion spo s when measu ing pa ially pola ized ligh . d DOP measu emen s o a se ies
o measu emen s whe e he DOP is cycled be ween 0 and 1 o e ou epe i ions. e Di e ence be ween wo
pola ime e s as a unc ion o DOP. Pola iza ion s a e measu emen s o pa ially pola ized ligh in a
Poinca e sphe e. The dis ance o he o igin is p opo ional o he DOP. Red and g een colo s espec i ely
indica e he measu emen s om he MS pola ime e be o e and a e calib a ion, while blue poin s a e
measu emen s by he e e ence pola ime e .
The calib a ion esul s in Figu es 4 and 5 we e done using ca e ully p epa ed pola iza ion s a es
and simul aneous measu emen s wi h bo h he MS and e e ence pola ime e s. To alida e he
MS pola ime e in a ealis ic use case i was es ed in a se up o digi al his opa hology35. The
calib a ion ma ices used a e he same as hose in Figu e 4, and he esul s a e p esen ed in
Figu e 6 oge he wi h a sepa a e e e ence measu emen done using he comme cial
pola ime e . Figu e 6 a is an image o a issue phan om consis ing o 2 ma e ials made o eplica e
he geome ical and op ical sca e ing p ope ies o a his ological issue block wi h a cance ous
inclusion38. Righ -handed ci cula ly pola ized ligh wi h DOP = 1 is ocused on o a spo on his
issue phan om and he sca e ed ligh is measu ed using he MS pola ime e . The spo is
scanned ac oss a egion o he issue phan om and Figu es 6 b, c a e maps showing how he DOP
o he sca e ed ligh a ies ac oss he sample as measu ed by he wo pola ime e s. As obse ed
in he DOP maps, bo h pola ime e s gi e simila esul s wi h lowe DOP alues in egions
mimicking he cance ous issue, al hough he MS pola ime e exhibi s compa a i ely mo e noise
in hese loca ions. To quan i y he di e ences be ween he wo pola ime e s, h ee a eas a e
selec ed and ma ked in blue, g een and magen a, espec i ely co esponding o a eas ou side
he inclusion, whe e he inclusion is hin and whe e i is hick – hese egions possess di e en
a e age DOP alues in he measu emen s. A b eakdown o all he pixel measu emen s wi hin
each o hese a eas is gi en in Figu es 6 d-h.
The main sou ce o di e ences be ween he wo pola ime e s can be unde s ood by looking a
he in ensi y dis ibu ions measu ed by he PAX1000 pola ime e shown in Figu e 6d. Ligh
sca e ed om he inclusion a eas ha e on a e age less in ensi y and a much highe in ensi y
a ia ion; in he magen a-colo ed a ea he in ensi y is e enly dis ibu ed ac oss 7 dBm, while o
he blue ma ked a ea ou side he inclusion mos measu emen s a e wi hin 2 dBm o each o he .
This, combined wi h he lowe DOP and co espondingly lowe con as be ween di ac ion spo
in ensi ies o he inclusion a eas esul in he no iceably la ge a ia ion in he measu emen s
om he MS pola ime e o hese a eas – look o example a he di e ence in dis ibu ion o he
magen a-colo ed measu emen s in Figu es 6 e and 6 . This di e ence is s ill isible when
b eaking down he pola iza ion s a es on componen basis in Figu es 6 g and 6 h, whe e he MS
pola ime e does well o he 𝑠𝑠1 componen bu is sligh ly o o 𝑠𝑠2 and 𝑠𝑠3 which ha e dis ibu ions
close o ze o. No ice also ha he MS pola ime e gi es a la ge DOP han he e e ence
pola ime e o he a eas ma ked wi h magen a and g een. This is a ibu ed o he ac ha he
calib a ion ma ices we e cons uc ed using ully pola ized ligh , and hus a e biased owa ds
highly pola ized ligh . Fu u e implemen a ions will mi iga e his e ec by including also he DOP
in he calib a ion as demons a ed in Figu e 5, bu o se e al di e en pola iza ion s a es sp ead
e enly ac oss he Poinca e sphe e a he han jus along one line.
In o al, he MS pola ime e does a good job cha ac e izing he pola iza ion s a es and he
di e en egions o he issue phan om a e easily dis inguished. Howe e , i was ound ha o
eal issue samples, which ha e less uni o mi y han he issue phan om and hus e en g ea e
a iabili y in sca e ed in ensi y and DOP, he cu en implemen a ion o he MS pola ime e is no
good enough in e ms o dynamic ange. While his can be pa ly mi iga ed by dynamically
adjus ing exposu e imes, he accompanying inc ease in acquisi ion ime when scanning he
whole sample is de imen al o eal use cases. A simple solu ion is o eplace he CMOS senso
wi h a s ip a ay o one pho odiode o each di ac ion o de . This modifica ion would inc ease
he dynamic ange and sensi i i y, as well as emo e he cu en bo leneck in scanning speed
which is limi ed by a maximum sampling a e o 400 Hz, de e mined by he speed o he o a ing
wa e pla e in he comme cial pola ime e . Fo e e ence he maximum ame a e o he MS
pola ime e is a ound 600 Hz, limi ed by he CMOS senso ame a e o he ele an subse o
pixels.
To conclude, we ha e designed, ab ica ed and cha ac e ized a dedica ed MS pola ime e ,
benchma king i agains a comme cial pola ime e o bo h ully and pa ially pola ized ligh o
wa eleng h 640 nm. We ha e ound ha bo h design and ab ica ion MS impe ec ions can be
calib a ed away using a simple p ocedu e in ol ing se e al calib a ion ma ices. Fu he mo e, we
ha e demons a ed i s use o digi al his opa hology, whe e i is expec ed o enable sys ems o
scanning issue samples ha a e as e , mo e compac and cheape han cu en labo a o y
p o o ypes. A eas o imp o emen include inc ease in he dynamic ange and mi iga ion o low
DOP measu emen s, bo h o which a e expec ed o be s aigh o wa dly implemen ed in u u e
de ices by swi ching o ewe bu mo e sensi i e pixels and imp o ing he se o pola iza ion
s a es used o calib a ion. We belie e ha , wi h hese imp o emen s in place, he conside ed
me asu ace pola ime e based de ices will be eady o p ac ical his opa hology applica ions in
clinical en i onmen .
Figu e 6. Benchma king wi h issue phan om. a Mic oscope image o a issue phan om used o
compa ing he MS pola ime e agains a e e ence pola ime e (Tho labs PAX1000). Wi hin he a ea ma ked
wi h an o ange squa e, he DOP is imaged poin wise using he e e ence pola ime e b and MS pola ime e
c. The measu emen poin s wi hin he a eas ma ked wi h blue, g een and magen a in b and c a e p esen ed
in mo e de ail in he emaining plo s d-h using he same colo s. d His og am o he eflec ed powe as
measu ed by he PAX1000. e and Pola iza ion s a es plo ed on a Poinca e sphe e wi h he DOP
ep esen ed as he adius, as measu ed by he PAX1000, e, and he MS pola ime e , . g and h His og ams
o he co esponding S okes pa ame e s and DOP plo ed o he PAX1000, g, and he MS pola ime e , h.
Colo ed iangles indica e he a e age alues o he a ious dis ibu ions.
Me hods
Me asu ace design and ab ica ion
The MS uni cells and supe cell g a ings we e simula ed and op imized using he fini e elemen
me hod (FEM) in COMSOL Mul iphysics 5.6. The pe mi i i y o Au was based on abula ed
alues39 bu wi h a ac o 3 inc ease in he imagina y pa o he pe mi i i y o Au ma e ial. This
adjus men gi es close ag eemen be ween simula ion and expe imen s likely due o su ace
oughness, g ain bounda y e ec s and inc eased damping om a i anium adhesion laye .
The MS was ab ica ed wi h elec on-beam li hog aphy (EBL) and a li -o p ocess: S a ing wi h
a chip om a polished Si wa e , a 3 nm Ti adhesion laye , a 120 nm Au laye , 3 nm Ti adhesion laye
and a 50 nm SiO2 laye we e deposi ed (To nado 400, C yo ox). This was ollowed by spin coa ing
o 100 nm PMMA (PMMA A2, Mic oChem). The MS pa e n was hen insc ibed using EBL (JEOL
JSM-6500F field-emission SEM wi h a Rai h Elphy Quan um li hog aphy sys em) and subsequen
de elopmen . The pa e n was con e ed o Au s uc u es by li -o in ace one a e deposi ion
o a 2 nm Ti adhesion laye and 50 nm Au laye (To nado 400, C yo ox).
Pola ime e ab ica ion and calib a ion
A model o he MS pola ime e was made in Zemax o op imize he choice o componen s and
hei alignmen . The pola ime e consis s o he MS subs a e glued o a non-pola izing beam
spli e (Tho labs BS010) wi h a plano-con ex lens (9 mm ocal leng h) glued on he opposi e side
as shown in Figu e 1, bo h using UV cu ing glue (No land Op ical Adhesi e 61). The beam spli e
was placed on a 6-axis s age connec ed o a CMOS came a (Tho labs CS165MU/M) using a
cus om b acke .
Alignmen and calib a ion we e done wi h he se up shown in Figu e 4. A 640 nm fibe coupled
lase was connec ed o a combined collima o /beam expande and sen h ough pola iza ion
con olling op ics. Pe ec alignmen o he beam on he MS was confi med by imaging he MS
su ace and beam posi ion h ough he beam spli e by empo a ily emo ing he in eg a ed
CMOS came a. To es he calib a ion o pa ially pola ized ligh , a lase was sen h ough wo
pola iza ion main aining fibe s, adjus ing he ela i e angle be ween hese wo fibe s changed he
DOP be ween 0% and 100%.
Tissue phan om measu emen s
A wo-componen phan om was designed o mimic a his ological issue block, combining
adipose and umo ous issues wi h complex geome ies. A eal his ological sample o bio issue
was used as a basis, wi h he ou lines o he umo ous issue ex ac ed om a pola ime ic scan
and con e ed o a 3D displacemen map. The phan om was subsequen ly 3D p in ed by
s e eoli hog aphy using UV-cu able esins (Fo mlabs Elas ic and Fo mlabs Clea ) wi h added zinc
oxide nanopa icles o model he sca e ing p ope ies o he adipose issue. Upon comple ion,
he p in ed objec was insed in a sol en ba h o emo e any emaining esin and hen cu ed in a
UV o en o ensu e ull solidifica ion and s uc u al ein o cemen . Finally, he cance ous
inclusion o he phan om was added using opaque esin (Pho ocen ic Flexible Whi e and Zo ax
Whi e I o y) and he su ace was polished a e cu ing in an ul a iole chambe 38.
The measu emen s o he issue phan om we e pe o med in a se up o pola ime ic
his opa hology35. A supe con inuum lase was fil e ed using an acous o-op ic unable fil e (640
nm le h ough) be o e going h ough s a ic pola iza ion con olling op ics and hen being ocused
on o he sample a an incidence angle o 55°. A 10× objec i e collec ed ligh a an angle o 30°,
which was collima ed in a 4F sys em using a 100 µm pinhole be o e en e ing he pola ime e . The
whole egion o in e es was mapped poin wise by ansla ing he sample be ween each
measu emen . Measu emen s using he wo pola ime e s we e done sepa a ely, hus he
mapped a eas a e sligh ly di e en o he wo and indi idual measu emen poin s canno be
compa ed di ec ly.
Re e ences
1. He, C. e al. Pola isa ion op ics o biomedical and clinical applica ions: a e iew. Ligh Sci Appl 10,
194 (2021).
2. Ghosh, N. Tissue pola ime y: concep s, challenges, applica ions, and ou look. J. Biomed. Op 16,
110801 (2011).
3. Ramella-Roman, J. C., Say ashe , I. & Piccini, M. A e iew o pola iza ion-based imaging echnologies
o clinical and p eclinical applica ions. J. Op . 22, 123001 (2020).
4. Qi, J. & Elson, D. S. Muelle pola ime ic imaging o su gical and diagnos ic applica ions: a e iew. J.
Biopho onics 10, 950–982 (2017).
5. Pie angelo, A. e al. Ex- i o cha ac e iza ion o human colon cance by Muelle pola ime ic imaging.
Op . Exp ess 19, 1582 (2011).
6. Ushenko, A. G. e al. Insigh s in o polyc ys alline mic os uc u e o blood films wi h 3D Muelle ma ix
imaging app oach. Sci Rep 14, 13679 (2024).
7. Qi, J. & Elson, D. S. A high defini ion Muelle pola ime ic endoscope o issue cha ac e isa ion. Sci
Rep 6, 25953 (2016).
8. Clancy, N. T. e al. Pola ised s e eo endoscope and na owband de ec ion o minimal access su ge y.
Biomed. Op . Exp ess 5, 4108 (2014).
9. Ushenko, V. A. e al. Embossed opog aphic depola isa ion maps o biological issues wi h di e en
mo phological s uc u es. Sci Rep 11, 3871 (2021).
10. Lope a, M. J., T usiak, M., Doblas, A., O e ae e, H. & T ujillo, C. Muelle -Gabo holog aphic
mic oscopy. Op ics and Lase s in Enginee ing 178, 108191 (2024).
11. Ushenko, V. A. e al. 3D Muelle ma ix mapping o laye ed dis ibu ions o depola isa ion deg ee
o analysis o p os a e adenoma and ca cinoma di use issues. Sci Rep 11, 5162 (2021).
12. Bo o ko a, M. e al. Sc eening o Alzheime ’s Disease Wi h Mul iwa eleng h S okes Pola ime y
in a Mouse Model. IEEE T ans. Med. Imaging 41, 977–982 (2022).
13. I ano , D. e al. Colon cance de ec ion by using Poinca é sphe e and 2D pola ime ic mapping o
ex i o colon samples. Jou nal o Biopho onics 13, e202000082 (2020).
14. Be y, H. G., Gab ielse, G. & Li ings on, A. E. Measu emen o he S okes pa ame e s o ligh .
Appl. Op . 16, 3200 (1977).
15. Po s, A., Nielsen, M. G. & Bozhe olnyi, S. I. Plasmonic me ag a ings o simul aneous
de e mina ion o S okes pa ame e s. Op ica 2, 716 (2015).
16. A babi, E., Kamali, S. M., A babi, A. & Fa aon, A. Full-S okes Imaging Pola ime y Using Dielec ic
Me asu aces. ACS Pho onics 5, 3132–3140 (2018).
17. Dong, J. & Zhou, H. Pola ime e s om bulky op ics o in eg a ed op ics: A e iew. Op ics
Communica ions 465, 125598 (2020).
18. Soma, G. e al. Compac and scalable pola ime ic sel -cohe en ecei e using a dielec ic
me asu ace. Op ica 10, 604 (2023).
19. Juhl, M., Muelle , J. P. B. & Leosson, K. Me asu ace Pola ime e on Op ical Fibe Face by Nano-
T ans e o UV-Cu able Hyb id Polyme . IEEE J. Selec . Topics Quan um Elec on. 25, 1–7 (2019).
20. Wu, W., Yu, Y., Liu, W. & Zhang, X. Fully in eg a ed CMOS-compa ible pola iza ion analyze .
Nanopho onics 8, 467–474 (2019).
21. Lee, K. e al. Ul acompac B oadband Plasmonic Pola ime e . Lase & Pho onics Re iews 12,
1700297 (2018).
22. Ren, Y. e al. A silicon me asu ace o ull-S okes pola ime y o in a ed ligh . Applied Physics
Le e s 122, 091103 (2023).
23. Shah, Y. D. e al. An All-Dielec ic Me asu ace Pola ime e . ACS Pho onics 9, 3245–3252 (2022).
24. Po s, A. & Bozhe olnyi, S. I. Wa eguide Me acouple s o In-Plane Pola ime y. Phys. Re . Applied
5, 064015 (2016).
25. Jung, M. e al. Pola ime y Using G aphene-In eg a ed Aniso opic Me asu aces. ACS Pho onics
5, 4283–4288 (2018).
26. Fang, C., Li, J., Zhou, B. & Li, D. Sel -powe ed Fil e less On-chip Full-S okes Pola ime e .
27. Ding, F., Po s, A., Chen, Y., Zenin, V. A. & Bozhe olnyi, S. I. Beam-Size-In a ian
Spec opola ime e s Using Gap-Plasmon Me asu aces. ACS Pho onics 4, 943–949 (2017).
28. Chen, W. T. e al. In eg a ed plasmonic me asu aces o spec opola ime y. Nano echnology
27, 224002 (2016).
29. Rubin, N. A. e al. Ma ix Fou ie op ics enables a compac ull-S okes pola iza ion came a.
Science 365, eaax1839 (2019).
30. Fan, Q. e al. Diso de ed me asu ace enabled single-sho ull-S okes pola iza ion imaging
le e aging weak dich oism. Na Commun 14, 7180 (2023).
31. Zuo, J. e al. Chip-in eg a ed me asu ace ull-S okes pola ime ic imaging senso . Ligh Sci Appl
12, 218 (2023).
32. Kuzne so , A. I. e al. Roadmap o Op ical Me asu aces. ACS Pho onics 11, 816–865 (2024).
33. Li, L. W., Rubin, N. A., Juhl, M., Pa k, J.-S. & Capasso, F. E alua ion and cha ac e iza ion o
imaging pola ime y h ough me asu ace pola iza ion g a ings. Appl. Op . 62, 1704 (2023).
34. Hsiao, H. e al. An Ul a-B oadband High E iciency Pola iza ion Beam Spli e o High Spec al
Resolu ion Pola ime ic Imaging in he Nea In a ed. Ad anced Science 9, 2201227 (2022).
35. I ano , D. e al. Pola iza ion-Based His opa hology Classifica ion o Ex Vi o Colon Samples
Suppo ed by Machine Lea ning. F on . Phys. 9, 814787 (2022).
36. Ding, F., Yang, Y., Deshpande, R. A. & Bozhe olnyi, S. I. A e iew o gap-su ace plasmon
me asu aces: undamen als and applica ions. Nanopho onics 7, 1129–1156 (2018).
37. Yang, Z. e al. Gene alized Ha mann-Shack a ay o dielec ic me alens sub-a ays o
pola ime ic beam p ofiling. Na Commun 9, 4607 (2018).
38. Lau i, J., A sie ich, T., Sie yi, O., Byko , A. & Fab i ius, T. 1.5-µm Op ical Cohe ence Tomog aphy
o Quali y Inspec ion o 3D-P in ed Sca e ing Phan oms. in 2024 IEEE In e na ional Ins umen a ion
and Measu emen Technology Con e ence (I2MTC) 1–5 (2024).
doi:10.1109/I2MTC60896.2024.10561156.
39. Johnson, P. B. & Ch is y, R. W. Op ical Cons an s o he Noble Me als. Phys. Re . B 6, 4370–4379
(1972).
Da a a ailabili y
The da a suppo ing he findings o his s udy a e a ailable om he co esponding au ho upon
easonable eques .
Acknowledgemen s
This esea ch has ecei ed unding om he ATTRACT p og am (Eu opean Union Ho izon 2020
Resea ch and Inno a ion P og am 101004462); Villum Fonden (37372, 50343, awa d in Technical
and Na u al Sciences 2019); Danma ks F ie Fo sknings ond (1134-00010B); No ges
Fo sknings åd (323322); Ho izon 2020 CA23125 - The mETama e ial oRmalism app oach o
ecognize cAnce (TETRA). This wo k was also pa ially suppo ed by he Resea ch Collabo a ions
g an (1203766815), unde he In e na ional Science Pa ne ships Fund unded by he UK
Depa men o Science Inno a ion and Technology in pa ne ship wi h he B i ish Council.
Au ho con ibu ions
P.T., C.M., C.A.D., S.I.B. and I.M. concei ed he idea. MS design and simula ion was done by C.M.,
while P.T. did he sys em in eg a ion design. C.M. ab ica ed he MS and bo h P.T. and C.M. did he
pola ime e in eg a ion. O.S. and A.B. ab ica ed he issue phan om. C.M. and O.S. pe o med
he measu emen s. P.T. a nd C. M . did he calib a ion and analyzed he esul s, wi h inpu om
S.I.B. and F.D. All au ho s con ibu ed o he discussion o he esul s and w i ing he manusc ip .
P.T. p o ided he fi s d a o he manusc ip . C.M., C.A.D. and S.I.B. supe ised he wo k.
Compe ing in e es s
The au ho s decla e no compe ing in e es s.
112 CHAPTER 5. ARTICLES
REFERENCES 119
78. Hu, Y. e al. 3D-In eg a ed me asu aces o ull-colou holog aphy. Ligh : Science
& Applica ions 8, 86. issn: 2047-7538. h ps://www.na u e.com/a icles/s41377-
019-0198-y (2025) (Sep . 2019).
79. Aksel od, G. M., Yang, Y. & Bowen, P. US11355858B2. h ps://pa en s.google.
com/pa en /US11355858B2/en (2025)(2022).
80. Yesilkoy, F. e al. Ul asensi i e hype spec al imaging and biode ec ion enabled by
dielec ic me asu aces. Na u e Pho onics 13, 390–396. issn: 1749-4885, 1749-4893.
h ps://www.na u e.com/a icles/s41566-019-0394-6 (2025) (June 2019).
81. Fa aji-Dana, M. e al. Compac olded me asu ace spec ome e . Na u e Communi-
ca ions 9, 4196. issn: 2041-1723. h ps://www.na u e.com/a icles/s41467-018-
06495-5 (2025) (Oc . 2018).
82. Jiang, H. e al. Me asu ace-enabled b oadband mul idimensional pho ode ec o s.
Na u e Communica ions 15, 8347. issn: 2041-1723. h ps : / / www . na u e . com /
a icles/s41467-024-52632-8 (2025) (Sep . 2024).
83. Wen, D. & C ozie , K. B. Semiconduc o lase s wi h in eg a ed me asu aces o di ec
ou pu beam modula ion, enabled by inno a i e ab ica ion me hods. Nanopho onics.
issn: 2192-8614. h ps://www.deg uy e .com/documen /doi/10.1515/nanoph-2022-
0585/h ml?lang=en (2023) (Jan. 2023).
84. Xie, Y.-Y. e al. Me asu ace-in eg a ed e ical ca i y su ace-emi ing lase s o
p og ammable di ec ional lasing emissions. Na u e Nano echnology 15, 125–130. issn:
1748-3387, 1748-3395. h ps://www.na u e. com / a icles / s41565 - 019 - 0611- y
(2025) (Feb. 2020).
85. S oo , H. e al. High-pu i y o bi al angula momen um s a es om a isible me asu -
ace lase . Na u e Pho onics 14, 498–503. issn: 1749-4885, 1749-4893. h p://www.
na u e.com/a icles/s41566-020-0623-z (2022) (Aug. 2020).
86. Kan, Y. e al. Me asu ace-Enabled Gene a ion o Ci cula ly Pola ized Single Pho-
ons. Ad anced Ma e ials 32, 1907832. issn: 0935-9648, 1521-4095. h ps://onlinelib a y.
wiley.com/doi/10.1002/adma.201907832 (2025) (Ap . 2020).
87. Sun, S. e al. G adien -index me a-su aces as a b idge linking p opaga ing wa es
and su ace wa es. Na u e Ma e ials 11, 426–431. issn: 1476-1122, 1476-4660. h ps:
//www.na u e.com/a icles/nma 3292 (2025) (May 2012).
88. Leng, B., Zhang, Y., Tsai, D. P. & Xiao, S. Me a-de ice: ad anced manu ac u ing.
Ligh : Ad anced Manu ac u ing 5, 117. issn: 2831-4093. h ps://www.ligh - am.
com/a icle/doi/10.37188/lam.2024.005 (2025) (2024).
89. Wang, Y. e al. High-e iciency b oadband ach oma ic me alens o nea -IR biological
imaging window. Na u e Communica ions 12, 5560. issn: 2041-1723. h ps://www.
na u e.com/a icles/s41467-021-25797-9 (2025) (Sep . 2021).
90. Zhu, X. e al. Resonan Lase P in ing o Op ical Me asu aces. Nano Le e s 22,
2786–2792. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/doi/10.1021/acs.
nanole .1c04874 (2025) (Ap . 2022).
91. McClung, A., To eh, M., Einck, V. J., Wa kins, J. J. & A babi, A. Visible Me alenses
wi h High Focusing E iciency Fab ica ed Using Nanoimp in Li hog aphy. Ad anced
Op ical Ma e ials 12, 2301865. issn: 2195-1071, 2195-1071. h ps://onlinelib a y.
wiley.com/doi/10.1002/adom.202301865 (2024) (Ma . 2024).
120 REFERENCES
92. Romano, L. e al. Me al assis ed chemical e ching o silicon in he gas phase: a
nano ab ica ion pla o m o X- ay op ics. Nanoscale Ho izons 5, 869–879. issn: 2055-
6756, 2055-6764. h ps://xlink. sc.o g/?DOI=C9NH00709A (2025) (2020).
93. Pa oux, A. e al. Challenges in nano ab ica ion o e icien op ical me asu aces.
Scien i ic Repo s 11, 5620. issn: 2045-2322. h ps://www.na u e.com/a icles/
s41598-021-84666-z (2024) (Ma . 2021).
94. Yang, Y. e al. The Road o Comme cializing Op ical Me asu aces: Cu en Chal-
lenges and Fu u e Di ec ions. ACS Nano 19, 3008–3018. issn: 1936-0851, 1936-086X.
h ps://pubs.acs.o g/doi/10.1021/acsnano.4c15474 (2025) (Jan. 2025).
95. Klop e , E. e al. Design o ab ica ion- ole an me a-a oms o pola iza ion-mul iplexed
me asu aces. MRS Communica ions 14, 990–999. issn: 2159-6867. h ps://link.
sp inge .com/10.1557/s43579-024-00629-1 (2025) (Sep . 2024).
96. Wang, E. W., Sell, D., Phan, T. & Fan, J. A. Robus design o opology-op imized
me asu aces. Op ical Ma e ials Exp ess 9, 469. issn: 2159-3930. h ps : / / opg .
op ica.o g/abs ac .c m?URI=ome-9-2-469 (2025) (Feb. 2019).
97. Malek, S. C., O e ig, A. C., Sh es ha, S. & Yu, N. Ac i e nonlocal me asu aces.
Nanopho onics 10, 655–665. issn: 2192-8614. h ps://www.deg uy e .com/documen /
doi/10.1515/nanoph-2020-0375/h ml (2022) (Jan. 2021).
98. O e ig, A. C., Malek, S. C., Ca e , M. J., Sh es ha, S. & Yu, N. Selec ion ules
o quasibound s a es in he con inuum. Physical Re iew B 102. issn: 2469-9950,
2469-9969. h ps://link.aps.o g/doi/10.1103/PhysRe B.102.035434 (2024) (July
2020).
99. Papagiakoumou, E., Ronzi i, E. & Emiliani, V. Scanless wo-pho on exci a ion wi h
empo al ocusing. Na u e Me hods 17, 571–581. issn: 1548-7091, 1548-7105. h p:
//www.na u e.com/a icles/s41592-020-0795-y (2022) (June 2020).
100. Johnson, P. B. & Ch is y, R. W. Op ical Cons an s o he Noble Me als. Physical
Re iew B 6, 4370–4379. issn: 0556-2805. h ps://link.aps.o g/doi/10.1103/
PhysRe B.6.4370 (2024) (Dec. 1972).
101. Meng, C., Th ane, P. C. V., Ding, F. & Bozhe olnyi, S. I. Full- ange bi e ingence
con ol wi h piezoelec ic MEMS-based me asu aces. Na u e Communica ions 13,
2071. issn: 2041-1723. h ps://doi.o g/10.1038/s41467-022-29798-0 (Ap . 2022).
102. Mandel, L. & Wol , E. Op ical Cohe ence and Quan um Op ics isbn: 978-0-521-
41711-2. h ps : / / www . camb idge . o g / co e / books / op ical - cohe ence - and -
quan um-op ics/F8CB94C70FA64CD3FB60890CA2048168 (2025) (Camb idge Uni e si y
P ess, Camb idge, 1995).
103. Do ah, A. H. & Capasso, F. Tunable s uc u ed ligh wi h la op ics. Science 376,
eabi6860. issn: 0036-8075, 1095-9203. h ps://www . science .o g/doi/10 . 1126 /
science.abi6860 (2022) (Ap . 2022).
104. Shal ou , A. M., Shalae , V. M. & B onge sma, M. L. Spa io empo al ligh con ol
wi h ac i e me asu aces. Science 364, eaa 3100. issn: 0036-8075, 1095-9203. h ps:
//www.science.o g/doi/10.1126/science.aa 3100 (2022) (May 2019).
105. Yang, J., Gu ung, S., Bej, S., Ni, P. & Howa d Lee, H. W. Ac i e op ical me asu aces:
comp ehensi e e iew on physics, mechanisms, and p ospec i e applica ions. Repo s
on P og ess in Physics 85, 036101. issn: 0034-4885, 1361-6633. h ps://iopscience.
iop.o g/a icle/10.1088/1361-6633/ac2aa (2024) (Ma . 2022).
REFERENCES 121
106. Yu, P., Li, J. & Liu, N. Elec ically Tunable Op ical Me asu aces o Dynamic Pola -
iza ion Con e sion. Nano Le e s 21, 6690–6695. issn: 1530-6984, 1530-6992. h ps:
//pubs.acs.o g/doi/10.1021/acs.nanole .1c02318 (2025) (Aug. 2021).
107. Decke , M. e al. Elec o-op ical swi ching by liquid-c ys al con olled me asu aces.
Op ics Exp ess 21, 8879. issn: 1094-4087. h ps://opg.op ica.o g/abs ac .c m?
URI=oe-21-7-8879 (2022) (Ap . 2013).
108. Sha ma, M., Tal, M., McDonnell, C. & Ellenbogen, T. Elec ically and all-op ically
swi chable nonlocal nonlinea me asu aces. Science Ad ances 9, eadh2353. issn:
2375-2548. h ps://www.science.o g/doi/10.1126/sciad .adh2353 (2025) (Aug.
2023).
109. Meye , C. e al. F ´eede icksz-Like T ansi ion in a Biaxial Smec ic- A Phase. Physical
Re iew X 11, 031012. issn: 2160-3308. h ps : / / link . aps . o g / doi / 10 . 1103 /
PhysRe X.11.031012 (2025) (July 2021).
110. Ding, F., Yang, Y. & Bozhe olnyi, S. I. Dynamic Me asu aces Using Phase-Change
Chalcogenides. Ad anced Op ical Ma e ials 7, 1801709. issn: 2195-1071, 2195-1071.
h ps://onlinelib a y.wiley.com/doi/10.1002/adom.201801709 (2025) (July
2019).
111. Abdollah amezani, S. e al. Elec ically d i en ep og ammable phase-change me a-
su ace eaching 80% e iciency. Na u e Communica ions 13, 1696. issn: 2041-1723.
h ps://www.na u e.com/a icles/s41467-022-29374-6 (2024) (Ma . 2022).
112. Li, B. e al. Fundamen al limi s o ansmission modula ion in VO2me asu aces.
Pho onics Resea ch 11, B40. issn: 2327-9125. h ps://opg.op ica.o g/abs ac .
c m?URI=p j-11-1-B40 (2025) (Jan. 2023).
113. Howes, A. e al. Op ical Limi ing Based on Huygens’ Me asu aces. Nano Le e s 20,
4638–4644. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/doi/10.1021/acs.
nanole .0c01574 (2025) (June 2020).
114. Jung, C., Lee, E. & Rho, J. The ise o elec ically unable me asu aces. Science
Ad ances 10, eado8964. issn: 2375-2548. h ps://www.science.o g/doi/10.1126/
sciad .ado8964 (2024) (Aug. 2024).
115. Damgaa d-Ca s ensen, C. Elec o-op ic me asu aces o ee-space ligh modula ion
PhD hesis (Syddansk Uni e si e , 2025). h ps://doi.o g/10.21996/74079b61-
a906-4 4a-9314-c a2 da 0598.
116. Guo, J. e al. Elec ically Tunable Gap Su ace Plasmon-based Me asu ace o Vis-
ible Ligh . Scien i ic Repo s 7, 14078. issn: 2045-2322. h ps://www.na u e.com/
a icles/s41598-017-14583-7 (2025) (Oc . 2017).
117. Yang, Y. e al. Fem osecond op ical pola iza ion swi ching using a cadmium oxide-
based pe ec abso be . Na u e Pho onics 11, 390–395. issn: 1749-4885, 1749-4893.
h ps://www.na u e.com/a icles/npho on.2017.64 (2025) (June 2017).
118. P yce, I. M., Aydin, K., Kelai a, Y. A., B iggs, R. M. & A wa e , H. A. Highly
S ained Complian Op ical Me ama e ials wi h La ge F equency Tunabili y. Nano
Le e s 10, 4222–4227. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/doi/10.
1021/nl102684x (2025) (Oc . 2010).
119. Ee, H.-S. & Aga wal, R. Tunable Me asu ace and Fla Op ical Zoom Lens on a
S e chable Subs a e. Nano Le e s 16, 2818–2823. issn: 1530-6984, 1530-6992. h ps:
//pubs.acs.o g/doi/10.1021/acs.nanole .6b00618 (2022) (Ap . 2016).
122 REFERENCES
120. Hou, X. e al. Bioinspi ed mul ichannel colo ul enc yp ion h ough ki igami ac i-
a ing g a ing. Science Bulle in 68, 276–283. issn: 20959273. h ps://linkinghub.
else ie .com/ e ie e/pii/S2095927323000440 (2025) (Feb. 2023).
121. Chen, C. e al. Tunable To oidal Response in a Recon igu able Te ahe z Me ama-
e ial. Ad anced Op ical Ma e ials 9, 2101215. issn: 2195-1071, 2195-1071. h ps:
//onlinelib a y.wiley.com/doi/10.1002/adom.202101215 (2025) (No . 2021).
122. Tao, H. e al. Recon igu able Te ahe z Me ama e ials. Physical Re iew Le e s 103,
147401. issn: 0031-9007, 1079-7114. h ps://link.aps.o g/doi/10.1103/PhysRe Le .
103.147401 (2025) (Oc . 2009).
123. Ou, J. Y., Plum, E., Jiang, L. & Zhelude , N. I. Recon igu able Pho onic Me ama-
e ials. Nano Le e s 11, 2142–2144. issn: 1530-6984, 1530-6992. h ps://pubs.acs.
o g/doi/10.1021/nl200791 (2025) (May 2011).
124. Mao, Y. e al. Mul i-Di ec ion-Tunable Th ee-Dimensional Me a-A oms o Re e sible
Swi ching be ween Midwa e and Long-Wa e In a ed Regimes. Nano Le e s 16,
7025–7029. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/doi/10.1021/acs.
nanole .6b03210 (2025) (No . 2016).
125. Zhao, Y. e al. The mosensi i e Plasmonic Colo Enabled by Sodium Me asu ace.
Ad anced Func ional Ma e ials 33, 2214492. issn: 1616-301X, 1616-3028. h ps :
//onlinelib a y.wiley.com/doi/10.1002/ad m.202214492 (2025) (Aug. 2023).
126. Chen, S., Chen, J., Zhang, X., Li, Z.-Y. & Li, J. Ki igami/o igami: un olding he
new egime o ad anced 3D mic o ab ica ion/nano ab ica ion wi h “ olding”. Ligh :
Science & Applica ions 9, 75. issn: 2047-7538. h ps://www.na u e.com/a icles/
s41377-020-0309-9 (2025) (Ap . 2020).
127. Chen, S. e al. Recon igu able nano-ki igami me asu aces by pneuma ic p essu e.
Pho onics Resea ch 8, 1177. issn: 2327-9125. h ps://opg.op ica.o g/abs ac .
c m?URI=p j-8-7-1177 (2025) (July 2020).
128. Zhao, Y. e al. Mechanically econ igu able me asu aces: ab ica ions and appli-
ca ions. npj Nanopho onics 1, 16. issn: 2948-216X. h ps : / / www . na u e . com /
a icles/s44310-024-00010-z (2025) (June 2024).
129. Ren, Z. e al. Le e aging o MEMS Technologies o Op ical Me ama e ials Appli-
ca ions. Ad anced Op ical Ma e ials 8, 1900653. issn: 2195-1071, 2195-1071. h ps:
//onlinelib a y.wiley.com/doi/10.1002/adom.201900653 (2025) (Feb. 2020).
130. Algamili, A. S. e al. A Re iew o Ac ua ion and Sensing Mechanisms in MEMS-
Based Senso De ices. Nanoscale Resea ch Le e s 16, 16. issn: 1556-276X. h ps:
//link.sp inge .com/10.1186/s11671-021-03481-7 (2025) (Jan. 2021).
131. Ahmad, M., Bah i, M. & Sawan, M. MEMS Mic omi o Ac ua ion Techniques:
A Comp ehensi e Re iew o T ends, Inno a ions, and Fu u e P ospec s. Mic oma-
chines 15, 1233. issn: 2072-666X. h ps://www.ncbi.nlm.nih.go /pmc/a icles/
PMC11509184/ (2025) (Sep . 2024).
132. Tao, H. e al. Mic owa e and e ahe z wa e sensing wi h me ama e ials. Op ics
Exp ess 19, 21620. issn: 1094-4087. h ps://opg.op ica.o g/oe/abs ac .c m?
u i=oe-19-22-21620 (2025) (Oc . 2011).
133. Bilgin, H., Zahe a , S., Sadeghzadeh, S., Yalcinkaya, A. D. & To un, H. A MEMS-
based e ahe z de ec o wi h me ama e ial-based abso be and op ical in e e ome -
ic eadou . Senso s and Ac ua o s A: Physical 244, 292–298. issn: 09244247. h ps:
//linkinghub.else ie .com/ e ie e/pii/S0924424716301662 (2025) (June 2016).
REFERENCES 123
134. Wang, H. e al. Using ligh o image millime e wa e based on s acked me a-MEMS
chip. Ligh : Science & Applica ions 14, 59. issn: 2047-7538. h ps://www.na u e.
com/a icles/s41377-024-01733-6 (2025) (Jan. 2025).
135. Han, Z., F ydendahl, C., Mazu ski, N. & Le y, U. MEMS can ile e –con olled plas-
monic colo s o sus ainable op ical displays. Science Ad ances 8, eabn0889. issn:
2375-2548. h ps://www.science.o g/doi/10.1126/sciad .abn0889 (2025) (Ap .
2022).
136. Hols een, A. L., Cihan, A. F. & B onge sma, M. L. Tempo al colo mixing and dy-
namic beam shaping wi h silicon me asu aces. Science 365, 257–260. issn: 0036-8075,
1095-9203. h ps://www.science.o g/doi/10.1126/science.aax5961 (2022) (July
2019).
137. Liu, X. & Padilla, W. J. Recon igu able oom empe a u e me ama e ial in a ed
emi e . Op ica 4, 430. issn: 2334-2536. h ps://opg.op ica.o g/abs ac .c m?
URI=op ica-4-4-430 (2025) (Ap . 2017).
138. P iboˇsek, J., Bainschab, M. & Sasaki, T. Va i ocal MEMS mi o s o high-speed axial
ocus scanning: a e iew. Mic osys ems & Nanoenginee ing 9, 135. issn: 2055-7434.
h ps://www.na u e.com/a icles/s41378-022-00481-0 (2024) (Oc . 2023).
139. Han, Z., Colbu n, S., Majumda , A. & B¨oh inge , K. F. MEMS-ac ua ed me asu ace
Al a ez lens. Mic osys ems & Nanoenginee ing 6, 79. issn: 2055-7434. h ps://www.
na u e.com/a icles/s41378-020-00190-6 (2025) (Oc . 2020).
140. Tang, H. e al. On-chip mul i-deg ee-o - eedom con ol o wo-dimensional ma e ials.
Na u e 632, 1038–1044. issn: 0028-0836, 1476-4687. h ps : / / www . na u e . com /
a icles/s41586-024-07826-x (2025) (Aug. 2024).
141. Roy, T. e al. Dynamic me asu ace lens based on MEMS echnology. APL Pho onics
3, 021302. issn: 2378-0967. h ps:// pubs . aip.o g/app / a icle /3/2/ 021302 /
122986/Dynamic-me asu ace-lens-based-on-MEMS- echnology (2025) (Feb. 2018).
142. Demi , K. & Unlu, M. Minia u e MEMS: No el Key Componen s Towa d Te a-
he z Recon igu abili y. Jou nal o Mic oelec omechanical Sys ems 29, 455–467.
issn: 1057-7157, 1941-0158. h ps : / / ieeexplo e . ieee . o g / documen / 9094643/
(2025) (Aug. 2020).
143. Xu, C., Ren, Z., Wei, J. & Lee, C. Recon igu able e ahe z me ama e ials: F om un-
damen al p inciples o ad anced 6G applica ions. iScience 25, 103799. issn: 25890042.
h ps://linkinghub.else ie .com/ e ie e/pii/S2589004222000694 (2025) (Feb.
2022).
144. Chen, S. e al. Elec omechanically econ igu able op ical nano-ki igami. Na u e Com-
munica ions 12, 1299. issn: 2041-1723. h ps://www.na u e.com/a icles/s41467-
021-21565-x (2025) (Feb. 2021).
145. Ki el, C. In oduc ion o solid s a e physics 8 h ed. isbn: 978-0-471-41526-8 (Wiley,
Hoboken, NJ, 2005).
146. Bakke, T. & Johansen, I.-R. PZT mic omi o wi h in eg a ed piezo esis i e posi-
ion senso s in 2012 In e na ional Con e ence on Op ical MEMS and Nanopho on-
ics (IEEE, Ban , AB, Canada, Aug. 2012), 192–193. isbn: 978-1-4577-1513-6. h p:
//ieeexplo e.ieee.o g/documen /6318868/ (2025).
124 REFERENCES
147. Bakke, T. e al. A no el ul a-plana , long-s oke and low- ol age piezoelec ic mi-
c omi o . Jou nal o Mic omechanics and Mic oenginee ing 20, 064010. issn: 0960-
1317, 1361-6439. h ps://iopscience.iop.o g/a icle/10.1088/0960-1317/20/6/
064010 (2022) (June 2010).
148. Meng, C. e al. Dynamic piezoelec ic MEMS-based op ical me asu aces. Science
Ad ances 7, eabg5639. issn: 2375-2548. h ps://www.science.o g/doi/10.1126/
sciad .abg5639 (2022) (June 2021).
149. Be khou , A. & Koende ink, A. F. Pe ec Abso p ion and Phase Singula i ies in
Plasmon An enna A ay E alons. ACS Pho onics 6, 2917–2925. issn: 2330-4022, 2330-
4022. h ps://pubs.acs.o g/doi/10.1021/acspho onics.9b01019 (2022) (No .
2019).
150. Be khou , A. & Koende ink, A. F. A simple ans e -ma ix model o me asu ace
mul ilaye sys ems. Nanopho onics 9, 3985–4007. issn: 2192-8614, 2192-8606. h ps:
//www. deg uy e . com / documen /doi/10.1515/nanoph- 2020- 0212/h ml (2022)
(Sep . 2020).
151. Th ane, P. C. V., Meng, C., Ding, F. & Bozhe olnyi, S. I. MEMS Tunable Me asu -
aces Based on Gap Plasmon o Fab y–P´e o Resonances. Nano Le e s 22, 6951–
6957. issn: 1530-6984. h ps://pubs.acs.o g/doi/abs/10.1021/acs.nanole .
2c01692 (2022) (Aug. 2022).
152. Di dal, C. A. e al. MEMS- unable dielec ic me asu ace lens using hin- ilm PZT o
la ge displacemen s a low ol ages. Op ics Le e s 47, 1049–1052. issn: 1539-4794.
h ps://opg.op ica.o g/ol/abs ac .c m?u i=ol-47-5-1049 (2022) (Ma . 2022).
153. A babi, E. e al. MEMS- unable dielec ic me asu ace lens. Na u e Communica ions
9, 812. issn: 2041-1723. h p://www.na u e.com/a icles/s41467- 018- 03155- 6
(2022) (Dec. 2018).
154. Deng, Y. e al. MEMS-in eg a ed me asu aces o dynamic linea pola ize s. Op ica
11, 326. issn: 2334-2536. h ps://opg.op ica.o g/abs ac .c m?URI=op ica-11-
3-326 (2024) (Ma . 2024).
155. Li, A. e al. Excep ional poin s and non-He mi ian pho onics a he nanoscale. Na u e
Nano echnology 18, 706–720. issn: 1748-3387, 1748-3395. h ps://www.na u e.com/
a icles/s41565-023-01408-0 (2025) (July 2023).
156. Wang, C. e al. MEMS-me asu ace–enabled mode-swi chable o ex lase s. Science
Ad ances 10, eadq6299. issn: 2375-2548. h ps://www.science.o g/doi/10.1126/
sciad .adq6299 (2024) (No . 2024).
157. Meng, C., Th ane, P. C. V., Wang, C., Ding, F. & Bozhe olnyi, S. I. MEMS- unable
opological bilaye me asu aces o econ igu able dual-s a e phase con ol. Op ica
11, 1556–1566. issn: 2334-2536. h ps://opg.op ica.o g/op ica/abs ac .c m?
u i=op ica-11-11-1556 (2024) (No . 2024).
158. Th ane, P. e al. Me asu ace Pola ime e o S uc u al Imaging and Tissue Diag-
nos ics Jan. 2025. h p://a xi .o g/abs/2501.05864 (2025).
159. I ano , D. e al. Pola iza ion-Based His opa hology Classi ica ion o Ex Vi o Colon
Samples Suppo ed by Machine Lea ning. F on ie s in Physics 9, 814787. issn: 2296-
424X. h ps://www. on ie sin.o g/a icles/10.3389/ phy.2021.814787/ ull
(2024) (Jan. 2022).
REFERENCES 125
160. Roy, K., Lee, J. E.-Y. & Lee, C. Thin- ilm PMUTs: a e iew o o e 40 yea s o
esea ch. Mic osys ems & Nanoenginee ing 9, 95. issn: 2055-7434. h ps://www.
na u e.com/a icles/s41378-023-00555-7 (2025) (July 2023).
161. Dahl-Hansen, R., Gjessing, J., Ma dilo ich, P., F agkiadakis, C. & Tho s ensen, J. Re-
liable Pb(Z ,Ti)O3-based hin ilm piezoelec ic mic omi o s o space-applica ions.
Applied Physics Le e s 121, 132901. issn: 0003-6951, 1077-3118. h ps://pubs.aip.
o g/apl/a icle/121/13/132901/2834262/Reliable- Pb-Z - Ti-O3-based- hin-
ilm-piezoelec ic (2025) (Sep . 2022).
162. Pa k, J. e al. All-solid-s a e spa ial ligh modula o wi h independen phase and
ampli ude con ol o h ee-dimensional LiDAR applica ions. Na u e Nano echnology
16, 69–76. issn: 1748-3387, 1748-3395. h ps://www.na u e.com/a icles/s41565-
020-00787-y (2022) (Jan. 2021).
163. Huang, Y.-W. e al. Ga e-Tunable Conduc ing Oxide Me asu aces. Nano Le e s 16,
5319–5325. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/doi/10.1021/acs.
nanole .6b00555 (2024) (Sep . 2016).
164. Sokhoyan, R. e al. Elec ically unable conduc ing oxide me asu aces o high powe
applica ions. Nanopho onics 12, 239–253. issn: 2192-8614. h ps://www.deg uy e .
com/documen /doi/10.1515/nanoph-2022-0594/h ml (2024) (Feb. 2023).
165. Thyaga ajan, K., Sokhoyan, R., Zo nbe g, L. & A wa e , H. A. Milli ol Modula ion o
Plasmonic Me asu ace Op ical Response ia Ionic Conduc ance. Ad anced Ma e ials
29, 1701044. issn: 0935-9648, 1521-4095. h ps://onlinelib a y.wiley.com/doi/
10.1002/adma.201701044 (2025) (Aug. 2017).
166. Geo ge, D. e al. Elec ically unable di ac ion e iciency om g a ings in Al-doped
ZnO. Applied Physics Le e s 110, 071110. issn: 0003-6951, 1077-3118. h ps: / /
pubs . aip . o g / apl / a icle / 110 / 7 / 071110 / 33872 / Elec ically - unable -
di ac ion-e iciency- om (2025) (Feb. 2017).
167. Chan, W. L. e al. A spa ial ligh modula o o e ahe z beams. Applied Physics Le -
e s 94, 213511. issn: 0003-6951, 1077-3118. h ps://pubs.aip.o g/apl/a icle/
94/21/213511/118804/A-spa ial-ligh -modula o - o - e ahe z-beams (2025)
(May 2009).
168. Pa k, J. e al. Dynamic he mal emission con ol wi h InAs-based plasmonic me a-
su aces. Science Ad ances 4, eaa 3163. issn: 2375-2548. h ps://www.science.o g/
doi/10.1126/sciad .aa 3163 (2025) (Dec. 2018).
169. Lee, J. e al. Ul a as Elec ically Tunable Pola i onic Me asu aces. Ad anced Op-
ical Ma e ials 2, 1057–1063. issn: 2195-1071, 2195-1071. h ps://onlinelib a y.
wiley.com/doi/10.1002/adom.201400185 (2025) (No . 2014).
170. Wu, P. C. e al. Dynamic beam s ee ing wi h all-dielec ic elec o-op ic III–V mul iple-
quan um-well me asu aces. Na u e Communica ions 10, 3654. issn: 2041-1723. h ps:
//www.na u e.com/a icles/s41467-019-11598-8 (2025) (Aug. 2019).
171. Li, Q. e al. A Pu cell-enabled monolaye semiconduc o ee-space op ical modula o .
Na u e Pho onics 17, 897–903. issn: 1749-4885, 1749-4893. h ps://www.na u e.
com/a icles/s41566-023-01250-9 (2025) (Oc . 2023).
172. Ande sen, T. I. e al. Beam s ee ing a he nanosecond ime scale wi h an a omically
hin e lec o . Na u e Communica ions 13, 3431. issn: 2041-1723. h ps : / / www .
na u e.com/a icles/s41467-022-29976-0 (2025) (June 2022).
126 REFERENCES
173. Van De G oep, J. e al. Exci on esonance uning o an a omically hin lens. Na u e
Pho onics 14, 426–430. issn: 1749-4885, 1749-4893. h ps : / / www . na u e . com /
a icles/s41566-020-0624-y (2025) (July 2020).
174. Yao, Y. e al. Elec ically Tunable Me asu ace Pe ec Abso be s o Ul a hin Mid-
In a ed Op ical Modula o s. Nano Le e s 14, 6526–6532. issn: 1530-6984, 1530-
6992. h ps://pubs.acs.o g/doi/10.1021/nl503104n (2025) (No . 2014).
175. Zeng, B. e al. Hyb id g aphene me asu aces o high-speed mid-in a ed ligh modu-
la ion and single-pixel imaging. Ligh : Science & Applica ions 7, 51. issn: 2047-7538.
h ps://www.na u e.com/a icles/s41377-018-0055-4 (2024) (Aug. 2018).
176. Zhang, X. e al. Dynamic Pho ochemical and Op oelec onic Con ol o Pho onic
Fano Resonances ia Monolaye MoS2T ions. Nano Le e s 18, 957–963. issn: 1530-
6984, 1530-6992. h ps: // pubs. acs. o g / doi / 10 . 1021 /acs .nanole .7b04355
(2025) (Feb. 2018).
177. Weiss, A. e al. Tunable Me asu ace Using Thin-Film Li hium Nioba e in he Telecom
Regime. ACS Pho onics 9, 605–612. issn: 2330-4022, 2330-4022. h ps://pubs.acs.
o g/doi/10.1021/acspho onics.1c01582 (2024) (Feb. 2022).
178. Weigand, H. e al. Enhanced Elec o-Op ic Modula ion in Resonan Me asu aces o
Li hium Nioba e. ACS Pho onics 8, 3004–3009. issn: 2330-4022, 2330-4022. h ps:
//pubs.acs.o g/doi/10.1021/acspho onics.1c00935 (2024) (Oc . 2021).
179. Benea-Chelmus, I.-C. e al. Gigahe z ee-space elec o-op ic modula o s based on
Mie esonances. Na u e Communica ions 13, 3170. issn: 2041-1723. h ps://www.
na u e.com/a icles/s41467-022-30451-z (2022) (June 2022).
180. Zhang, L. e al. Plasmonic me a ibe s elec o-op ic modula o s. Ligh : Science &
Applica ions 12, 198. issn: 2047-7538. h ps://www.na u e.com/a icles/s41377-
023-01255-7 (2024) (Aug. 2023).
181. Damgaa d-Ca s ensen, C., Thomaschewski, M. & Bozhe olnyi, S. I. Elec o-op ic
me asu ace-based ee-space modula o s. Nanoscale 14, 11407–11414. issn: 2040-
3364, 2040-3372. h p://xlink. sc.o g/?DOI=D2NR02979K (2022) (2022).
182. Damgaa d-Ca s ensen, C. & Bozhe olnyi, S. I. Nonlocal elec o-op ic me asu aces
o ee-space ligh modula ion. Nanopho onics 12, 2953–2962. issn: 2192-8606, 2192-
8614. h ps://www.deg uy e .com/documen /doi/10.1515/nanoph-2023-0042/h ml
(2025) (July 2023).
183. Pi o a, S. e al. Fas ampli ude modula ion up o 1.5 GHz o mid-IR ee-space beams
a oom- empe a u e. Na u e Communica ions 12, 799. issn: 2041-1723. h ps://
www.na u e.com/a icles/s41467-020-20710-2 (2025) (Feb. 2021).
184. Sokhoyan, R., Hail, C. U., Foley, M., G ajowe , M. Y. & A wa e , H. A. All-Dielec ic
High-Q Dynamically Tunable T ansmissi e Me asu aces. Lase & Pho onics Re iews
18, 2300980. issn: 1863-8880, 1863-8899. h ps://onlinelib a y.wiley.com/doi/
10.1002/lpo .202300980 (2025) (June 2024).
185. Fang, Z. e al. Ul a-low-ene gy p og ammable non- ola ile silicon pho onics based on
phase-change ma e ials wi h g aphene hea e s. Na u e Nano echnology 17, 842–848.
issn: 1748-3387, 1748-3395. h ps://www.na u e.com/a icles/s41565-022-01153-
w(2024) (Aug. 2022).
REFERENCES 127
186. Michel, A.-K. U. e al. Re e sible Op ical Swi ching o In a ed An enna Resonances
wi h Ul a hin Phase-Change Laye s Using Fem osecond Lase Pulses. ACS Pho onics
1, 833–839. issn: 2330-4022, 2330-4022. h ps : / / pubs . acs . o g / doi / 10 . 1021 /
ph500121d (2025) (Sep . 2014).
187. Kim, Y. e al. Phase Modula ion wi h Elec ically Tunable Vanadium Dioxide Phase-
Change Me asu aces. Nano Le e s 19, 3961–3968. issn: 1530-6984, 1530-6992. h ps:
//pubs.acs.o g/doi/10.1021/acs.nanole .9b01246 (2025) (June 2019).
188. Chen, R. e al. Non- ola ile elec ically p og ammable in eg a ed pho onics wi h a
5-bi ope a ion. Na u e Communica ions 14, 3465. issn: 2041-1723. h ps://www.
na u e.com/a icles/s41467-023-39180-3 (2024) (June 2023).
189. Li, J., Yu, P., Zhang, S. & Liu, N. Elec ically-con olled digi al me asu ace de ice o
ligh p ojec ion displays. Na u e Communica ions 11, 3574. issn: 2041-1723. h ps:
//www.na u e.com/a icles/s41467-020-17390-3 (2024) (July 2020).
190. Moi a, P. e al. Elec ically Tunable Re lec i e Me asu aces wi h Con inuous and
Full-Phase Modula ion o High-E iciency Wa e on Con ol a Visible F equencies.
ACS Nano 17, 16952–16959. issn: 1936-0851, 1936-086X. h ps://pubs.acs.o g/
doi/10.1021/acsnano.3c04071 (2025) (Sep . 2023).
191. Kowe dziej, R., W ´obel, J. & Kula, P. Ul a as elec ical swi ching o nanos uc u ed
me ade ice wi h dual- equency liquid c ys al. Scien i ic Repo s 9, 20367. issn: 2045-
2322. h ps://www.na u e.com/a icles/s41598-019-55656-z (2024) (Dec. 2019).
192. Li, K. e al. Elec ically Swi chable, Pola iza ion-Sensi i e Enc yp ion Based on Alu-
minum Nanoape u e A ays In eg a ed wi h Polyme -Dispe sed Liquid C ys als.
Nano Le e s 21, 7183–7190. issn: 1530-6984, 1530-6992. h ps://pubs.acs.o g/
doi/10.1021/acs.nanole .1c01947 (2024) (Sep . 2021).
193. Kossy e , P. A. e al. Elec ic Field Tuning o Plasmonic Response o Nanodo A ay
in Liquid C ys al Ma ix. Nano Le e s 5, 1978–1981. issn: 1530-6984, 1530-6992.
h ps://pubs.acs.o g/doi/10.1021/nl0513535 (2025) (Oc . 2005).
194. Wu, X. Y. e al. An Ul a hin, Fas -Response, La ge-Scale Liquid-C ys al-Facili a ed
Mul i-Func ional Recon igu able Me asu ace o Comp ehensi e Wa e on Mod-
ula ion. Ad anced Ma e ials 36, 2402170. issn: 0935-9648, 1521-4095. h ps : / /
onlinelib a y.wiley.com/doi/10.1002/adma.202402170 (2025) (June 2024).
195. Chang, X. e al. Fas -swi ching econ igu able me ade ice wi h me asu ace-induced
liquid c ys al alignmen o ligh modula o applica ions. Op ical Ma e ials Exp ess
14, 1094. issn: 2159-3930. h ps://opg.op ica.o g/abs ac .c m?URI=ome-14-4-
1094 (2025) (Ap . 2024).
196. Ou, J.-Y., Plum, E., Zhang, J. & Zhelude , N. An elec omechanically econ ig-
u able plasmonic me ama e ial ope a ing in he nea -in a ed. Na u e nano echnology
8(Ma . 2013).
197. Zhang, X., Kwon, K., Hen iksson, J., Luo, J. & Wu, M. C. A la ge-scale mic oelec omechanical-
sys ems-based silicon pho onics LiDAR. Na u e 603, 253–258. issn: 1476-4687. h ps:
//www.na u e.com/a icles/s41586-022-04415-8 (2022) (Ma . 2022).
198. Kim, D. U. e al. P og ammable pho onic a ays based on mic oelec omechanical
elemen s wi h em owa -le el s andby powe consump ion. Na u e Pho onics 17,
1089–1096. issn: 1749-4885, 1749-4893. h ps://www.na u e.com/a icles/s41566-
023-01327-5 (2024) (Dec. 2023).
128 REFERENCES
199. Zheng, T., Kwon, H. & Fa aon, A. Nanoelec omechanical Tuning o High- QSlo
Me asu aces. Nano Le e s 23, 5588–5594. issn: 1530-6984, 1530-6992. h ps : / /
pubs.acs.o g/doi/10.1021/acs.nanole .3c00999 (2024) (June 2023).
200. Zhu, W. M. e al. Swi chable Magne ic Me ama e ials Using Mic omachining P o-
cesses. Ad anced Ma e ials 23, 1792–1796. issn: 0935-9648, 1521-4095. h ps : / /
onlinelib a y.wiley.com/doi/10.1002/adma.201004341 (2025) (Ap . 2011).
201. Ho ie, Y., A babi, A., A babi, E., Kamali, S. M. & Fa aon, A. High-Speed, Phase-
Dominan Spa ial Ligh Modula ion wi h Silicon-Based Ac i e Resonan An ennas.
ACS Pho onics 5, 1711–1717. issn: 2330-4022, 2330-4022. h ps://pubs.acs.o g/
doi/10.1021/acspho onics.7b01073 (2025) (May 2018).
202. Zhu, W. e al. Mic oelec omechanical Mal ese-c oss me ama e ial wi h unable e -
ahe z aniso opy. Na u e Communica ions 3, 1274. issn: 2041-1723. h ps://www.
na u e.com/a icles/ncomms2285 (2025) (Dec. 2012).
