Supplementay Figue Supplementay Figue 1 Optical etup chematic diagam. See Supplementay Method fo detailed deciption.
Supplementay Figue 2 lectical ignal flow diagam. A function geneato (AFG1) geneate two 50 MHz continuou wave ignal epaately fom two channel to dive the AOM. A fou-phae tepping digital hologaphy method i ued to captue the field of the light catteed fom the ample. Duing DOPC ecoding, a data acquiition cad (DAQ) geneate a voltage tepping ignal to modulate the phae of the ignal fom channel 2 of AFG1. At the ame time, it output a ynchonized tigge ignal (35 Hz) to the camea of the DOPC ytem. Afte the fit field i captued, the DAQ geneate a gating ignal to enable the ultaound tone but fom the econd function geneato (AFG2). The phae ecoding poce take ~260 m and the playback take ~20 m, leading to a total of ~280 m fo the TRUM poce. Abbeviation: AFG Abitay Function geneato; AMP Amplifie; AOM Acouto-Optic Modulato; Ch Channel; DAQ Data Acquiition Cad; PM Phae modulation; UST Ultaound Tanduce.
Supplementay Figue 3 xpeimentally meaued modulation efficiency of diect ultaound modulation and micobubble modulation. (a) Ultaound modulation efficiency baed on the hologaphical image of the ultaound focu in clea wate. (b) Meaued micobubble modulation efficiency baed on the field ubtaction method of TRUM. The mean modulation efficiency M wa calculated by aveaging the modulation efficiency on each pixel ove the ROI (cicle), which i defined by the FWHM of the modulation efficiency aco cente of each taget in the hoizontal diection.
Supplementay Figue 4 Meauing the pobability ditibution of micobubble detuction ove peue. A micobubble heet wa placed at the focal plane of an ultaound beam. We diectly obeved the detuction of micobubble with a camea and counted the numbe of detucted micobubble at diffeent peue. (a) Ditibution of detucted micobubble ove peue. The poition of the detucted micobubble wee extacted uing a watehed algoithm (ee Method). The big cicle indicate the FWHM diamete of the ultaound beam (~85 µm). We counted only the micobubble within the mall cicle (40 µm), within which the peue wa appoximated to be unifom. (b) Dak field image of the micobubble befoe and afte inonation with a equence of peue. (c) Reulting pobability denity function (PDF) and cumulative ditibution function (CDF) of the micobubble detuction ove peue. Scale ba: 50 µm.
Supplementay Figue 5 xploing TRUM nonlineaity with 20 MHz ultaound beam. Only lateal eolution wa demontated becaue the axial focal zone of the 20 MHz ultaound beam (2.3 mm, -6dB) exceed the field of view of ou obevation ytem. We ecued the micobubble in agaoe gel in an acylic tube following the ame potocol a decibed in the Method. In ode to demontate the elationhip between lateal confinement of the TRUM foci and ultaound peue, we implemented TRUM to the ame ample with 10 level of peue (linealy fom 0.6 to 2.8 MPa) in acending ode. 91 et of data wee captued by tageting at diffeent poition of the ample. The TRUM foci wee captued and poceed in the ame way a decibed the main aticle. (a) Image of TRUM foci at diffeent ultaound peue goup fom a typical et of data. (b) Coeponding micobubble image befoe and afte micobubble detuction. (c) TRUM focu poition ditibution baed on 91 et of data, including that hown in (a). (d) Compaion of the hitogam pofile (FWHM of the fitted Gauian pofile: 95 µm) of the lowe peue goup and the ultaound lateal peue pofile (FWHM: 210 µm). Scale ba 50 µm.
Supplementay Figue 6 Raw image of the TRUM nonlineaity expeiment (Figue 5b in the main aticle) and the coeponding optical image of the ultaound beam. (a) xample image of the TRUM foci, which i one of the 135 et of image that contibute to Figue 5b in the main aticle. In thi cae, we claified the image captued with 15 peue level into 3 peue goup a indicated on the top of each figue and then aveaged the image within the ame goup. (b) The coeponding micobubble image befoe and afte inonation with a equence of peue. (c) The optical image of the ultaound beam obtained by uing digital hologaphy. In thi cae, we imaged the ultaound beam to the obevation camea in the obevation module (Supplementay Fig. 1) and acquied the optical field infomation by intefeing with an exta efeence beam (not hown in Supplementay Fig. 1). Scale ba: 50 µm. Supplementay Figue 7 Optical focuing on ize deceaed micobubble. (a) Image of the micobubble befoe applying ultaound. (b) Afte applying ultaound. (c) Light focuing on the hunk micobubble. (d) The focu vanihed a the SLM phae patten hifted by 10 pixel in both hoizontal and vetical diection. Scale ba: 10 µm.
Supplementay Note 1 Focuing on ize-deceaed micobubble. The mechanim of micobubble detuction include fagmentation and diffuion. While the fagmentation occu within the timecale of micoecond, the acoutically diven diffuion effect take up to ten of milliecond 1. Fundamentally, lowe ultaound peue and lage micobubble diamete eult in a longe time fo the ga diolution in the uounding medium. In ou expeiment, we found that acoutically diven diffuion of micobubble ga take longe in gel than in aline (PBS). Ideally fo TRUM, the micobubble ae detoyed within the ultaound duation (~29 m in ou expeiment) a hown in Supplementay Movie 1 (cale ba: 10 µm). Howeve, we alo obeved incomplete detuction of micobubble within that peiod (Supplementay Movie 2), which typically occu with low ultaound peue and lage micobubble ize. Nevethele, thi effect alo enable an optical focu (Supplementay Fig. 7) to be fomed. Altenatively, a longe inonation time allow the micobubble ga to completely diolute into the uounding medium (Supplementay Movie 3). Supplementay Method Optical etup The optical etup chematic diagam i hown in Supplementay Fig. 1. The etup can be patitioned into fou majo module: light ouce egulation, DOPC, ample opeation, and obevation. The light ouce egulation module povide thee lae beam: a ample beam, efeence beam, and playback beam. The intenity of each lae beam i egulated by a waveplate and a polaizing beam plitte. Both the ample beam and the efeence beam ae fequencyhifted by 50 MHz uing acouto-optic modulato (AOM). Duing hologam ecoding, the ignal to AOM1 i phae-hifted following the tigge ignal to the camea of the DOPC ytem. A path length matching am i applied to the efeence beam, a ou lae ha a elative hot coheent length (~7 mm). All beam ae patially filteed uing ingle mode fibe and collimated fo DOPC application. The polaization of all beam i et to the hoizontal diection, matching the polaization of the SLM. Duing DOPC ecoding, the hutte in the playback beam path (SH1) i cloed and hutte in the ample beam path (SH2) i open. Both hutte flip afte ecoding. The DOPC module conit of a CMOS camea (CAMS) and a SLM. Thee two component ae peciely aligned 2 though a plate beam plitte (BSP) o that the optical field ecoded by the camea can be econtucted faithfully in pace by the SLM. A polaize i ued to match the polaization of the ample beam and SLM. The ample beam and efeence beam ae combined though a 90% beam plitte (BST). The ecoded phae of the ample beam i conjugated and ent to the SLM, which modulate the collimated playback beam. A beam compenato i ued to compenate fo the wavefont ditotion intoduced by the plate beam plitte. In the ample opeation module, light tanmitted though the ample i collected by a 50 mm len (L4) whoe focu i poitioned in the middle of the ample. An apetue (AP3) i
ued to contol the peckle aea, which cove ~9 pixel in ou cae. A tage poitioned ultaound tanduce i placed above the ample (nomal to the diagam) but it i flattened to the left hand ide in thi two-dimenional diagam. The mio (M4) fo fluoecence ignal detection i placed below the ample but alo flattened in thi diagam. The obevation module ha an imaging ytem and a fluoecence detection ytem. In the imaging ytem, a 20 objective (OB) and a 100 mm tube len (L6) ae aanged in a 4-f configuation though which the micobubble in the ample ae imaged onto a camea (CAM). In the expeiment that demontate cytomety, the fluoecence emitted fom the ample i collected by a len (L5) and filteed by a fluoecence filte (FF). A ingle photon counting avalanche photodiode (APD) i ued to detect the fluoecence ignal. lectical ignal flow fo TRU The TRU optical focu hown in Figue 2e wa ceated uing the ame etup but with a modified ignal flow (not hown in Supplementay Fig. 2). Intead of modulating the ample beam with an AOM, TRU ued ingle cycle ultaound to hift the fequency of the ample beam o that the ultaound encoded light wa meaued. The efeence beam wa modulated at 50.010 MHz athe than 50 MHz o that the camea aveaged out the patten due to intefeence between unmodulated light and efeence beam, which wa othewie locked by the 20 khz lae pule. In thi cae, each ultaound pule wa phae-inveed fom the peceding pule o that only ultaound modulated light i locked by the efeence beam 3. The field meauement fo TRU wa the ame a that in TRUM. PBR meauement PBR wa calculated fom a egion of inteet (ROI, 600 pixel (x) by 40 pixel (y)) cented on the tube fo both TRUM and TRU foci. Fit, a one dimenional focu pofile wa extacted fom the ROI fo peak calculation. Fo the TRUM focu, the ow aco the cente of the focu wa ued a the one-dimenional focu. Fo the TRU focu (Figue 2e), the onedimenional focu wa calculated by aveaging the ROI in the y diection. In the econd tep, a one-dimenional Gauian pofile wa fitted to each one-dimenional focu pofile. The amplitude of the fitted Gauian pofile wa conideed a the peak intenity. Finally, to calculate the backgound light intenity, the SLM wa hifted by 10 pixel in both x and y diection afte the focu wa made and the backgound image (e.g. Figue 3d in the main aticle) captued. The backgound intenity wa etimated by aveaging the ROI on the backgound image. The PBR wa then calculated by taking the atio between peak intenity and backgound intenity. Modulation efficiency meauement To compae the modulation efficiency of guideta ued in TRU and TRUM, we meaued the light field modulated by the ultaound focu and micobubble in clea media epectively and calculated the modulation efficiency. The field image of the guideta wee captued uing the camea of the DOPC ytem. a. Ultaound modulation We meaued the ultaound modulated light field baed on the lock-in cheme and 4-phae tepping hologaphy ued fo the phae ecoding of a TRU poce (ee lectical ignal
flow). Mathematically, the optical field on the eno plane can be decompoed into a efeence field ( ) k and a ignal field modulated field m and an unmodulated field tepping method. The complex field ( k ),, which can be futhe decompoed into a u. k denote the tep numbe of the phae m and u ae given by ( k) A exp i 0 a t k 2 m Am exp i 0 a t m u Au exp i0t u, (1) whee A and denote the amplitude and phae of each complex field with aociated ubcipt, 0 and a ae the fequency of the light and ultaound epectively. It hould be noted that the ultaound i aumed to be continuou wave hee fo implicity. The light intenity on the eno plane at the k th tepping phae can then be expeed a 2 2 I( k) ( k) ( k). (2) m u By ubtituting the field tem with the ight ide of Supplementay quation (1), we can futhe expand Supplementay quation (2) to I( k) A A A 2A A co k 2 2 2 2 m u m m 2A A co t k 2 2A A co t k 2. u a u m u a m u (3) The AC tem ( k) and ha an ocillation fequency of * u m * u a. A the expoue time of the eno i much longe than the ultaound peiod ( 2 ), thee ae effectively aveaged out to 0. Then, the intenity captued fom the eno can be appoximated a 2 2 2 m u m m I( k) A A A 2A A co k 2. (4) With the fou meaued intefeence patten (at k 0, 1, 2 and 3 ), we ae able to compute the modulated field which i optically amplified by the efeence light field: a 4A A expi I I i I I c 0 2 1 3 m, (5) whee the elative phae diffeence eplace m. Then, by auming the ideal modulation ( Am A), the maximum poible amplitude of the computed field can be calculated: Finally, we define the modulation efficiency Max 4A A. (6) c
2 2 c c M Max( ) 16I I c, (7) whee the I and I ae the efeence ( 2 A ) and ignal beam intenity ( 2 A ) epectively. I and I wee meaued epaately to calculate the modulation efficiency. The modulation efficiency i then aveaged ove the ROI, whoe diamete i the full width at half maximum (FWHM) of the one-dimenional modulation efficiency pofile in the hoizontal diection (Supplementay Fig. 3). In thi expeiment, the ultaound peak peue i ~2 MPa, and the calculated modulation efficiency i ~0.5%. b. Micobubble modulation We meaued the micobubble modulated light field with phae tepping hologaphy and field ubtaction method ued fo the phae ecoding of a TRUM poce. The efeence field ( ) k, ignal field befoe modulation expeed a and ignal field afte modulation ( k) A exp i 0t k 2 am A exp i0t A exp i 0t can be, (8) whee the pime ( ) ymbol denote the ignal field befoe modulation. Hee we aume that the micobubble modulate the amplitude and phae of ignal field by a m (anging fom 0 to 1) and epectively. Then, the light intenity at the th k tepping phae i given by befoe : I ( k) ( k) afte : I( k) ( k) 2. (9) By ubtituting the field tem with the ight ide of Supplementay quation (8), we have 2 2 2 befoe : I ( k) A am A 2am A A co k 2 2 2 afte : I( k) A A 2A A co k 2. (10) With the fou meaued intefeence patten fo each field meauement, we compute the ignal field (amplified by the efeence field): befoe: c 4am A A exp i. afte: c 4A A exp i (11) Following the field ubtaction method ued in the TRUM poce, we ae able to calculate the amplitude of the modulated field c (amplified by the efeence field): c c m c 4A A 1 a exp i. (12)
Maximum poible amplitude i achieved when am 1 and : Max 8A A, (13) c c which i twice of the maximum poible amplitude fo TRU poce (ee Supplementay quation (6)). Thi oiginate fom the natue of field ubtaction method involving two light field meauement. Finally, we define the modulation efficiency M 2 2 c c c c Max( ) c 64II c. (14) We again meaued the efeence beam intenity I and ignal beam intenity I epaately and calculated the modulation efficiency. Simila to the ultaound modulation cae, the modulation efficiency i aveaged ove the ROI a hown Supplementay Fig. 3. The diamete of the ROI i again the FHWM of the one-dimenional modulation pofile in the hoizontal diection. Micobubble detuction pobability ditibution To chaacteie the micobubble detuction pobability ove peue, we diectly counted the numbe of detucted micobubble at 31 dicete peue level (linealy fom 0.15 to 6.88 MPa). It hould be noted that TRUM wa not implemented hee, a the goal of thi method wa to count the numbe of micobubble detuction event. In thi cae, we deigned a micobubble heet (ee Method) and placed it at the focal plane of the ultaound beam. In ode to paallelize the meauement, we ued an ultaound beam with a wide peue pofile, which wa geneated by a lowe numeic-apetue tanduce (12.7 mm element diamete, 12.7 mm focal length, 50 MHz nominal cental fequency, PI-50, Olympu). The tanduce wa diven by an ultaound but ignal that had the ame fequency (45 MHz), numbe of cycle (10) and inteval time (10 µ) a the ignal fo the V3330 tanduce. An image of the micobubble heet wa captued afte each inonation to the micobubble, eulting in a et of 31 image at a ingle location. We epeated thi poce 95 time by tageting a new egion and captued 95 et of data. A watehed baed algoithm (ee Method) wa again ued to extact the poition of the detucted micobubble a hown in Supplementay Fig. 4a, and all the data et wee accumulated and claified into thee peue goup. The lage cicle (in magenta) outline the FWHM contou of the peue (diamete: ~85 µm), but we counted only thoe within the mall cicle (diamete: 40 µm), whee the peue wa appoximately unifom. Supplementay Fig. 4b how typical image of the micobubble befoe and afte inonation with a equence of peue. Contay to the micobubble image hown in the main aticle, hee the image wee captued in dak field becaue the ultaound beam wa nomal to the micobubble heet and oblique illumination wa ued. Supplementay Fig. 4c how the pobability denity function (PDF) and cumulative ditibution function (CDF) of the micobubble detuction ove peue.
Supplementay Refeence 1. Choma, J.., Dayton, P., Allen, J., Mogan, K. & Feaa, K. W. Mechanim of contat agent detuction. I Tan. Ultaon. Feoelect. Feq. Contol 48, 232 48 (2001). 2. Jang, M., Ruan, H., Zhou, H., Judkewitz, B. & Yang, C. Method fo auto-alignment of digital optical phae conjugation ytem baed on digital popagation. Opt. xpe 22, 14054 71 (2014). 3. Ruan, H., Mathe, M. L. & Mogan, S. P. Puled ultaound modulated optical tomogaphy with hamonic lock-in hologaphy detection. J. Opt. Soc. Am. A. Opt. Image Sci. Vi. 30, 1409 16 (2013).