One-step variation included compact modeling with conditional variational autoencoder

•A machine learning assisted compact modeling approach is proposed to include variation in one step by conditional variational autoencoder.•Simplicity and high accuracy are achieved and statistical results generated by the model can reflect the actual situation.•The trained model is compatible with...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Solid-state electronics Ročník 227; s. 109119
Hlavní autoři: Wang, Shuhan, Zhou, Zheng, Tang, Zili, Xu, Jinghan, Liu, Xiaoyan, Zhang, Xing
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 01.08.2025
Témata:
ISSN:0038-1101
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract •A machine learning assisted compact modeling approach is proposed to include variation in one step by conditional variational autoencoder.•Simplicity and high accuracy are achieved and statistical results generated by the model can reflect the actual situation.•The trained model is compatible with SPICE, thus can simulate variation in circuits. Efficient and accurate variation modeling serves as a critical part in circuit evaluation, which can reproduce actual electrical behavior of semiconductor devices. Conventional variation modeling usually consists of two steps: compact modeling the basic electrical properties and sub-modeling the variation sources introduced in MOSFET manufacturing process, mostly structural and doping parameters. This lengthy process results in a gap between device production and rapid circuit analysis. In order to improve modeling efficiency, in this work, we propose a one-step variation-included compact modeling approach leveraging machine learning. Utilizing conditional variational autoencoder (cVAE), currents with variation are directly constructed without the sub-modeling step, as variation sources of the cVAE model are automatically extracted. Benchmark against prior distribution of dataset generated by Monte Carlo simulation of BSIM-CMG, normalized variation in figure of merits (FoMs) of cVAE generated I-V curves are all above 0.9. After implementing the model in SPICE, high accuracy in circuit-level variation modeling also indicates the potential of the proposed model.
AbstractList •A machine learning assisted compact modeling approach is proposed to include variation in one step by conditional variational autoencoder.•Simplicity and high accuracy are achieved and statistical results generated by the model can reflect the actual situation.•The trained model is compatible with SPICE, thus can simulate variation in circuits. Efficient and accurate variation modeling serves as a critical part in circuit evaluation, which can reproduce actual electrical behavior of semiconductor devices. Conventional variation modeling usually consists of two steps: compact modeling the basic electrical properties and sub-modeling the variation sources introduced in MOSFET manufacturing process, mostly structural and doping parameters. This lengthy process results in a gap between device production and rapid circuit analysis. In order to improve modeling efficiency, in this work, we propose a one-step variation-included compact modeling approach leveraging machine learning. Utilizing conditional variational autoencoder (cVAE), currents with variation are directly constructed without the sub-modeling step, as variation sources of the cVAE model are automatically extracted. Benchmark against prior distribution of dataset generated by Monte Carlo simulation of BSIM-CMG, normalized variation in figure of merits (FoMs) of cVAE generated I-V curves are all above 0.9. After implementing the model in SPICE, high accuracy in circuit-level variation modeling also indicates the potential of the proposed model.
ArticleNumber 109119
Author Tang, Zili
Xu, Jinghan
Liu, Xiaoyan
Zhou, Zheng
Zhang, Xing
Wang, Shuhan
Author_xml – sequence: 1
  givenname: Shuhan
  surname: Wang
  fullname: Wang, Shuhan
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
– sequence: 2
  givenname: Zheng
  surname: Zhou
  fullname: Zhou, Zheng
  email: zhouzime@pku.edu.cn
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
– sequence: 3
  givenname: Zili
  surname: Tang
  fullname: Tang, Zili
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
– sequence: 4
  givenname: Jinghan
  surname: Xu
  fullname: Xu, Jinghan
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
– sequence: 5
  givenname: Xiaoyan
  surname: Liu
  fullname: Liu, Xiaoyan
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
– sequence: 6
  givenname: Xing
  surname: Zhang
  fullname: Zhang, Xing
  email: zhx@pku.edu.cn
  organization: School of Integrated Circuits, Peking University, Beijing 100871, China
BookMark eNp9kMtKAzEUhrOoYFt9AHfzAlNPJsk4wZUUL4VCN3YdchtNmSYlSSu-vRlGcOfq3P7v8PMv0MwHbxG6w7DCgNv7wyolu2qgYWXmGPMZmgOQrsblfI0WKR0AoGkxzNF-522dsj1VFxmdzC74ynk9nI01lQ7Hk9S5OgZjB-c_qi-XP8vWGzcK5fAHlV6ec7BeF228QVe9HJK9_a1LtH95fl-_1dvd62b9tK11w3CudUsU6xjoVpEOeGcIfeCcSEp7hqFRXBNtGtq3CjhVwKiRrFe6w7KnXFFOlghPf3UMKUXbi1N0Rxm_BQYxZiEOomQhxizElEVhHifGFmMXZ6NI2hXf1rhodRYmuH_oH4JVbGA
Cites_doi 10.1109/TED.2023.3274510
10.1016/j.sse.2007.04.008
10.1145/3670474.3685972
10.3390/electronics11172761
10.1109/JEDS.2020.3024669
10.1109/TED.2020.3048918
10.1109/TCAD.2023.3285032
10.1016/j.microrel.2008.07.039
10.1109/TCAD.2010.2042892
10.1007/978-3-7091-6827-1_73
10.1109/TCAD.2022.3193330
10.1109/ACCESS.2020.3014470
10.1109/TED.2018.2790083
ContentType Journal Article
Copyright 2025 Elsevier Ltd
Copyright_xml – notice: 2025 Elsevier Ltd
DBID AAYXX
CITATION
DOI 10.1016/j.sse.2025.109119
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
ExternalDocumentID 10_1016_j_sse_2025_109119
S0038110125000644
GroupedDBID --K
--M
-~X
.DC
.~1
0R~
123
1B1
1RT
1~.
1~5
4.4
457
4G.
5VS
6TJ
7-5
71M
8P~
9JN
AABXZ
AAEDT
AAEDW
AAEPC
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AATTM
AAXKI
AAXUO
AAYWO
ABDPE
ABFNM
ABFRF
ABJNI
ABMAC
ABNEU
ABWVN
ABXDB
ABXRA
ACDAQ
ACFVG
ACGFO
ACGFS
ACNCT
ACNNM
ACRLP
ACRPL
ACVFH
ADBBV
ADCNI
ADEZE
ADMUD
ADNMO
ADTZH
AEBSH
AECPX
AEFWE
AEIPS
AEKER
AENEX
AEUPX
AEZYN
AFFNX
AFJKZ
AFPUW
AFRZQ
AFTJW
AFXIZ
AGCQF
AGHFR
AGQPQ
AGRNS
AGUBO
AGYEJ
AHHHB
AHJVU
AIEXJ
AIGII
AIIUN
AIKHN
AITUG
AIVDX
AKBMS
AKRWK
AKYEP
ALMA_UNASSIGNED_HOLDINGS
AMRAJ
ANKPU
APXCP
ASPBG
AVWKF
AXJTR
AZFZN
BBWZM
BJAXD
BKOJK
BLXMC
BNPGV
CS3
DU5
EBS
EFJIC
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-2
G-Q
GBLVA
HMV
HVGLF
HZ~
H~9
IHE
J1W
JJJVA
KOM
LY7
M24
M38
M41
MAGPM
MO0
N9A
NDZJH
O-L
O9-
OAUVE
OGIMB
OZT
P-8
P-9
P2P
PC.
PZZ
Q38
R2-
RIG
RNS
ROL
RPZ
SDF
SDG
SDP
SES
SET
SEW
SMS
SPC
SPCBC
SPD
SPG
SSH
SSM
SSQ
SST
SSZ
T5K
TAE
TN5
WH7
WUQ
XSW
ZMT
ZY4
~G-
9DU
AAYXX
ACLOT
CITATION
EFKBS
EFLBG
~HD
ID FETCH-LOGICAL-c251t-c63b5850c6b38098d347993a44f5102b9c3cd24f6b094b054da5fbc81af49b493
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001464940600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0038-1101
IngestDate Sat Nov 29 07:55:41 EST 2025
Sat Jun 21 16:54:45 EDT 2025
IsPeerReviewed true
IsScholarly true
Keywords Variational autoencoder
Variation modeling
Compact model
Machine learning
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c251t-c63b5850c6b38098d347993a44f5102b9c3cd24f6b094b054da5fbc81af49b493
ParticipantIDs crossref_primary_10_1016_j_sse_2025_109119
elsevier_sciencedirect_doi_10_1016_j_sse_2025_109119
PublicationCentury 2000
PublicationDate August 2025
2025-08-00
PublicationDateYYYYMMDD 2025-08-01
PublicationDate_xml – month: 08
  year: 2025
  text: August 2025
PublicationDecade 2020
PublicationTitle Solid-state electronics
PublicationYear 2025
Publisher Elsevier Ltd
Publisher_xml – name: Elsevier Ltd
References Alam (b0005) 2008; 48
Wei, Wang, Zhao, Jiang, Wan (b0030) 2022; 42
Yang, Sun, Li, Shi, Liu (b0080) 2023; 70
Kingma DP. Auto-encoding variational bayes
Li (b0020) 2024
Sohn, Lee, Yan (b0085) 2015; 28
Choe, Kwak, Yu (b0025) 2023
Mehta, Raju, Xiao, Wang, Zhang, Wong (b0055) 2020; 8
Woo, Jeong, Choi, Cho, Kong, Kim (b0040) 2022; 11
Zhang, Jiang, Wang, Guo, Wang, Huang (b0075) 2018; 65
Jan. 1998.
Katto (b0100) 2007; 51
Li, McAndrew, Wu, Chaudhry, Victory, Gildenblat (b0045) 2010; 29
Wong (b0060) 2020; 8
Dai (b0065) 2023; 42
Lin S, Roberts SJ, Trigoni N, Clark R, Balancing reconstruction quality and regularisation in ELBO for VAEs; Sep. 2019
[Online]. Available: https://arxiv.org/abs/1909.03765.
Jiang, Wang, Wang, Cheng, Asenov, Huang (b0070) 2015
Dietrich, Haase (b0010) 2011
Wang, Kim, Ryu, Jeong, Choi, Kim (b0035) 2021; 68
2013.
Zhang L, Peng B, Li Y, Liu H, Dai W, Wang R. When device modeling meets machine learning: opportunities and challenges
Sep. 2024.
C. C. McAndrew, Statistical Circuit Modeling
10.1016/j.sse.2025.109119_b0090
Wang (10.1016/j.sse.2025.109119_b0035) 2021; 68
Wong (10.1016/j.sse.2025.109119_b0060) 2020; 8
Dai (10.1016/j.sse.2025.109119_b0065) 2023; 42
Woo (10.1016/j.sse.2025.109119_b0040) 2022; 11
Yang (10.1016/j.sse.2025.109119_b0080) 2023; 70
Choe (10.1016/j.sse.2025.109119_b0025) 2023
Katto (10.1016/j.sse.2025.109119_b0100) 2007; 51
Wei (10.1016/j.sse.2025.109119_b0030) 2022; 42
Sohn (10.1016/j.sse.2025.109119_b0085) 2015; 28
Mehta (10.1016/j.sse.2025.109119_b0055) 2020; 8
Jiang (10.1016/j.sse.2025.109119_b0070) 2015
Alam (10.1016/j.sse.2025.109119_b0005) 2008; 48
10.1016/j.sse.2025.109119_b0015
Li (10.1016/j.sse.2025.109119_b0045) 2010; 29
Zhang (10.1016/j.sse.2025.109119_b0075) 2018; 65
Li (10.1016/j.sse.2025.109119_b0020) 2024
10.1016/j.sse.2025.109119_b0050
Dietrich (10.1016/j.sse.2025.109119_b0010) 2011
10.1016/j.sse.2025.109119_b0095
References_xml – reference: Zhang L, Peng B, Li Y, Liu H, Dai W, Wang R. When device modeling meets machine learning: opportunities and challenges,
– volume: 8
  start-page: 143519
  year: 2020
  end-page: 143529
  ident: b0055
  article-title: Improvement of TCAD augmented machine learning using autoencoder for semiconductor variation identification and inverse design
– volume: 68
  start-page: 1318
  year: 2021
  end-page: 1325
  ident: b0035
  article-title: Artificial neural network-based compact modeling methodology for advanced transistors
– reference: . [Online]. Available: https://arxiv.org/abs/1909.03765.
– volume: 48
  start-page: 1114
  year: 2008
  end-page: 1122
  ident: b0005
  article-title: Reliability- and process-variation aware design of integrated circuits
  publication-title: ,
– volume: 8
  start-page: 992
  year: 2020
  end-page: 1000
  ident: b0060
  article-title: TCAD-machine learning framework for device variation and operating temperature analysis with experimental demonstration
  publication-title: ,
– volume: 70
  start-page: 3935
  year: 2023
  end-page: 3942
  ident: b0080
  article-title: Compact modeling of process variations in nanosheet complementary FET (CFET) and circuit performance predictions
– volume: 51
  start-page: 905
  year: 2007
  end-page: 912
  ident: b0100
  article-title: MOSFET I–V characteristics at small and large drain biases in the linear region
– reference: ; 2013.
– volume: 11
  start-page: 2761
  year: 2022
  ident: b0040
  article-title: Machine-learning-based compact modeling for sub-3-nm-node emerging transistors
– year: 2024
  ident: b0020
  article-title: Overview of emerging semiconductor device model methodologies: from device physics to machine learning engines
– volume: 42
  start-page: 5156
  year: 2023
  end-page: 5160
  ident: b0065
  article-title: Statistical compact modeling with artificial neural networks
– year: 2023
  ident: b0025
  article-title: Machine learning-assisted compact modeling of w-doped indium oxide channel transistor for back-end-of-line applications
– reference: Kingma DP. Auto-encoding variational bayes,
– volume: 42
  start-page: 1250
  year: 2022
  end-page: 1254
  ident: b0030
  article-title: A new compact MOSFET model based on artificial neural network with unique data preprocessing and sampling techniques
– reference: , Jan. 1998.
– volume: 65
  start-page: 847
  year: 2018
  end-page: 854
  ident: b0075
  article-title: Extraction of process variation parameters in FinFET technology based on compact modeling and characterization
– year: 2011
  ident: b0010
– reference: C. C. McAndrew, Statistical Circuit Modeling,
– volume: 28
  year: 2015
  ident: b0085
  article-title: Learning structured output representation using deep conditional generative models
– reference: Lin S, Roberts SJ, Trigoni N, Clark R, Balancing reconstruction quality and regularisation in ELBO for VAEs; Sep. 2019,
– reference: , Sep. 2024.
– volume: 29
  start-page: 599
  year: 2010
  end-page: 606
  ident: b0045
  article-title: Statistical modeling with the PSP MOSFET model
– year: 2015
  ident: b0070
  article-title: Predictive compact modeling of random variations in FinFET technology for 16/14nm node and beyond
– year: 2024
  ident: 10.1016/j.sse.2025.109119_b0020
  article-title: Overview of emerging semiconductor device model methodologies: from device physics to machine learning engines
  publication-title: Fundam Res
– ident: 10.1016/j.sse.2025.109119_b0090
– volume: 70
  start-page: 3935
  issue: 7
  year: 2023
  ident: 10.1016/j.sse.2025.109119_b0080
  article-title: Compact modeling of process variations in nanosheet complementary FET (CFET) and circuit performance predictions
  publication-title: IEEE Trans Electron Devices
  doi: 10.1109/TED.2023.3274510
– volume: 51
  start-page: 905
  issue: 6
  year: 2007
  ident: 10.1016/j.sse.2025.109119_b0100
  article-title: MOSFET I–V characteristics at small and large drain biases in the linear region
  publication-title: Solid-State Electron
  doi: 10.1016/j.sse.2007.04.008
– ident: 10.1016/j.sse.2025.109119_b0015
  doi: 10.1145/3670474.3685972
– volume: 11
  start-page: 2761
  issue: 17
  year: 2022
  ident: 10.1016/j.sse.2025.109119_b0040
  article-title: Machine-learning-based compact modeling for sub-3-nm-node emerging transistors
  publication-title: Electronics
  doi: 10.3390/electronics11172761
– volume: 8
  start-page: 992
  year: 2020
  ident: 10.1016/j.sse.2025.109119_b0060
  article-title: TCAD-machine learning framework for device variation and operating temperature analysis with experimental demonstration
  publication-title: IEEE J Electron Devices Soc,
  doi: 10.1109/JEDS.2020.3024669
– ident: 10.1016/j.sse.2025.109119_b0095
– volume: 68
  start-page: 1318
  issue: 3
  year: 2021
  ident: 10.1016/j.sse.2025.109119_b0035
  article-title: Artificial neural network-based compact modeling methodology for advanced transistors
  publication-title: IEEE Trans Electron Devices
  doi: 10.1109/TED.2020.3048918
– volume: 28
  year: 2015
  ident: 10.1016/j.sse.2025.109119_b0085
  article-title: Learning structured output representation using deep conditional generative models
  publication-title: Adv Neural Inf Proces Syst NIPS
– volume: 42
  start-page: 5156
  issue: 12
  year: 2023
  ident: 10.1016/j.sse.2025.109119_b0065
  article-title: Statistical compact modeling with artificial neural networks
  publication-title: IEEE Trans. Comput.-Aided Design Integr. Circuits Syst.
  doi: 10.1109/TCAD.2023.3285032
– volume: 48
  start-page: 1114
  issue: 8
  year: 2008
  ident: 10.1016/j.sse.2025.109119_b0005
  article-title: Reliability- and process-variation aware design of integrated circuits
  publication-title: Microelectron Reliab,
  doi: 10.1016/j.microrel.2008.07.039
– volume: 29
  start-page: 599
  issue: 4
  year: 2010
  ident: 10.1016/j.sse.2025.109119_b0045
  article-title: Statistical modeling with the PSP MOSFET model
  publication-title: IEEE Trans. Comput.-Aided Design Integr. Circuits Syst.
  doi: 10.1109/TCAD.2010.2042892
– year: 2023
  ident: 10.1016/j.sse.2025.109119_b0025
  article-title: Machine learning-assisted compact modeling of w-doped indium oxide channel transistor for back-end-of-line applications
  publication-title: IEEE Trans Electron Devices
– ident: 10.1016/j.sse.2025.109119_b0050
  doi: 10.1007/978-3-7091-6827-1_73
– year: 2011
  ident: 10.1016/j.sse.2025.109119_b0010
– volume: 42
  start-page: 1250
  issue: 4
  year: 2022
  ident: 10.1016/j.sse.2025.109119_b0030
  article-title: A new compact MOSFET model based on artificial neural network with unique data preprocessing and sampling techniques
  publication-title: IEEE Trans. Comput.-Aided Design Integr. Circuits Syst.
  doi: 10.1109/TCAD.2022.3193330
– volume: 8
  start-page: 143519
  year: 2020
  ident: 10.1016/j.sse.2025.109119_b0055
  article-title: Improvement of TCAD augmented machine learning using autoencoder for semiconductor variation identification and inverse design
  publication-title: IEEE Access
  doi: 10.1109/ACCESS.2020.3014470
– year: 2015
  ident: 10.1016/j.sse.2025.109119_b0070
  article-title: Predictive compact modeling of random variations in FinFET technology for 16/14nm node and beyond
  publication-title: IEEE Int Electron Devices Meet (IEDM)
– volume: 65
  start-page: 847
  issue: 3
  year: 2018
  ident: 10.1016/j.sse.2025.109119_b0075
  article-title: Extraction of process variation parameters in FinFET technology based on compact modeling and characterization
  publication-title: IEEE Trans Electron Devices
  doi: 10.1109/TED.2018.2790083
SSID ssj0002610
Score 2.4355562
Snippet •A machine learning assisted compact modeling approach is proposed to include variation in one step by conditional variational autoencoder.•Simplicity and high...
SourceID crossref
elsevier
SourceType Index Database
Publisher
StartPage 109119
SubjectTerms Compact model
Machine learning
Variation modeling
Variational autoencoder
Title One-step variation included compact modeling with conditional variational autoencoder
URI https://dx.doi.org/10.1016/j.sse.2025.109119
Volume 227
WOSCitedRecordID wos001464940600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  issn: 0038-1101
  databaseCode: AIEXJ
  dateStart: 19950101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: https://www.sciencedirect.com
  omitProxy: false
  ssIdentifier: ssj0002610
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3PT9swFLY64DAOCBgI2EA-7AQyamKnjY8IMfFLbNKAVVyi2E7UVFVa0aTiz9_zrzQbTBpIXKLIitvofdbL957f-4zQV02JezKnpCvCPmEqEARuOAmzQHRVAATBHNN5f92_uYkHA_6j07nyvTDzcb8s46cnPn1XqGEMwNats6-Au_lRGIB7AB2uADtc_wv472VGALrp0RzC4NTVMspxrTJlK85lZc-_adKwEBKrwuUEm0laQ6CuJlrnUrkK3pHv6B0XiphOpKPFKToNN__lM9DDerhYeg_DSW02QoaZ-1aadIF99KEYF35sYB67hJfzs11OIoyaijiXKPPNMovKJON8wbcC2wjazje0ygDPHLnNKYyOZzOtZRpGWvYqcM71T33sn2azU8uURYZgsQ9oOexHHFzc8snF2eCy-TBDsOhUOu17-E1uU-731x-9TFNa1ON2Ha25mAGfWKw3UCcrN9FqS0nyE7rzqOMGQOxRxw517FHHGnXcQh23UMct1LfQ3bez29Nz4k7MIBJ4akVkjwqI_7qyJ2jc5bHSfcKcpozl4HtDwSWVKmR5T0BUL4CtqzTKhYyDNGdcME630VI5KbMdhGNdwJhRobQmpKARlzlnUgkWUZZyHu2iQ2-gZGqFURJfMThKwJqJtmZirbmLmDdh4pidZWwJ4P3vaXtvm_YZfVwsyi9oqXqss320IudVMXs8cKviN18rc9Y
linkProvider Elsevier
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=One-step+variation+included+compact+modeling+with+conditional+variational+autoencoder&rft.jtitle=Solid-state+electronics&rft.au=Wang%2C+Shuhan&rft.au=Zhou%2C+Zheng&rft.au=Tang%2C+Zili&rft.au=Xu%2C+Jinghan&rft.date=2025-08-01&rft.pub=Elsevier+Ltd&rft.issn=0038-1101&rft.volume=227&rft_id=info:doi/10.1016%2Fj.sse.2025.109119&rft.externalDocID=S0038110125000644
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0038-1101&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0038-1101&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0038-1101&client=summon