Skip to main content

Technische Universität Dortmund

Patent Portfolio in Transition — the university’s patent families 2013–2024, compared across two complete five-year filing cohorts (2013–2017 and 2018–2022).

Data: EPO PATSTAT Global, Spring 2026 (patstat_2026a) Period: 2013–2024 Updated: 15 August 2026 Author: mtc.berlin
123
Patent Families
2013–2024, by first filing year
247
Applications
filed by TU Dortmund
46%
PCT Rate
57 families with a WO member
16
Filing Authorities
reached by the 123 families
21
Co-Filed Families
17% of 123 families

Executive Summary

What the patent record of Technische Universität Dortmund shows for filing years 2013 to 2024.

Between 2013 and 2024, Technische Universität Dortmund is named as applicant on 247 patent applications belonging to 123 DOCDB patent families. Each family is assigned to a single year — the earliest filing year across all of its members — so the annual figures add up to the total without double counting. The portfolio is modest in size and remarkably steady: annual output moves between 7 and 15 families, with no year outside that band.

The report compares two complete five-year filing cohorts, 2013–2017 (55 families) and 2018–2022 (46 families). Both end in a filing year that PATSTAT Spring 2026 records completely. The 22 families first filed in 2023 and 2024 appear in the annual trend and in the list of recent titles, but they are held out of every cohort comparison, because filings from those years are still being published and any cohort ending there would measure publication lag rather than filing behaviour.

Across the two cohorts the shape of the portfolio changes more than its size. Protection abroad becomes slightly more common: the share of families with a PCT (WO) member rises from 43.6 % to 47.8 % and the share reaching the United States from 14.5 % to 21.7 %, while the share with a European (EP) member falls from 60.0 % to 54.3 %. The average number of patent authorities per family rises from 2.05 to 2.24. At the same time the share of families that never leave the German office grows from 27.3 % to 37.0 %, so the portfolio is pulling apart into a more international and a purely national half.

The technology mix moves as well. Measured in WIPO technology sectors, Mechanical Engineering drops from 26 to 13 families while Instruments rises from 13 to 15, and at IPC subclass level the sharpest movement is sheet-metal working (B21D, 16 families to 3) against materials investigation and analysis (G01N, 4 families to 9). The inventor record moves in step: the most frequently named inventor, A. Erman Tekkaya, is on 18 of the 55 families in the first cohort and 6 of the 46 in the second, while 8 of the 24 inventors named on three or more families have no family at all in the first cohort.

Collaboration remains the exception rather than the rule. 21 of 123 families (17.1 %) carry a second applicant, spread over 21 partner organisations and four individuals; only two partners — Universität Duisburg-Essen and the Max-Planck-Gesellschaft — appear in more than one period.

Over twelve filing years the portfolio holds its volume and changes its content. The data shows a mechanical-engineering core that shrinks by half between the two cohorts and a measurement and sensing profile that grows, together with a wider international footprint per family. With 7 to 15 families per year, every figure in this report rests on small numbers; the report therefore states absolute counts alongside every percentage.

Filing Activity 2013–2024

Patent families by first filing year and the applications TU Dortmund filed for them. Each family is counted once, in the earliest filing year of any of its members.

Data table: Families and applications by first filing year
First filing year Patent families Applications Cohort
20139192013–2017
201411252013–2017
20157142013–2017
201615262013–2017
201713282013–2017
20187142018–2022
201912252018–2022
202010192018–2022
202110222018–2022
20227192018–2022
2023*1119held out
2024*1117held out
Total123247

* 2023 and 2024 are provisional: applications from these filing years are still being published and added with each PATSTAT release (EPO PATSTAT Global, Spring 2026). They are shown here but excluded from every cohort comparison in this report.

Annual output stays inside a narrow band: between 7 and 15 families in every one of the twelve years, with an average of 11.0 families per year in 2013–2017 and 9.2 in 2018–2022. Applications per family are close to constant as well (2.04 in the first cohort, 2.15 in the second). With single-digit to low-double-digit counts, a difference of two or three families in one year is not a trend; the cohort view below is the more reliable comparison.

Cohort Comparison: 2013–2017 vs. 2018–2022

Two complete five-year filing cohorts, side by side. Families are assigned to a cohort by their earliest filing year, so no family appears in both.

Data table: Cohort comparison
Metric 2013–2017 2018–2022
Patent families5546
Applications by TU Dortmund11299
Applications per family2.042.15
Families with a DE member41 (74.5 %)36 (78.3 %)
Families with an EP member33 (60.0 %)25 (54.3 %)
Families with a WO (PCT) member — PCT rate24 (43.6 %)22 (47.8 %)
Families with a US member8 (14.5 %)10 (21.7 %)
Families filed at DE only15 (27.3 %)17 (37.0 %)
Families at a single authority20 (36.4 %)19 (41.3 %)
Co-filed families (second applicant)11 (20.0 %)7 (15.2 %)
Average patent authorities per family2.052.24
Distinct patent authorities reached1012
Families with at least one granted member*30 (54.5 %)10 (21.7 %)

* Grant figures are not comparable between the cohorts: applications filed from 2018 onwards are still under examination at every office. See the Grant Outcomes section for the pendency caveat. All coverage figures count family members, i.e. a family counts once for an authority no matter how many applications it has there.

The two cohorts are close in size (55 and 46 families) and differ in reach. The PCT rate rises from 43.6 % to 47.8 % and the US share from 14.5 % to 21.7 %, while the EP share falls from 60.0 % to 54.3 %. Average authorities per family rise from 2.05 to 2.24. At the same time more families stay at the German office alone (15 to 17 families, 27.3 % to 37.0 %). This is consistent with a two-track portfolio: fewer, but more broadly protected international families alongside a growing set of national-only filings. The underlying differences are 2 to 7 families, so the direction is better supported than the magnitude.

Technology Profile — WIPO Sectors and Fields

Families mapped to the WIPO technology classification (35 fields in 5 sectors) via the IPC codes of all family members.

Data table: WIPO sectors by cohort
WIPO sector 2013–2017 2018–2022 2023–2024* Change (families)
Mechanical engineering26135−13
Chemistry18176−1
Electrical engineering17166−1
Instruments13157+2
Other fields103−1
Sum of sectors756127
Distinct families554622
Data table: WIPO technology fields with 5 or more families
WIPO field Sector 2013–2017 2018–2022 2023–2024* Total
MeasurementInstruments812626
Machine toolsMechanical engineering177125
Other special machinesMechanical engineering84214
TelecommunicationsElectrical engineering46212
Electrical machinery, apparatus, energyElectrical engineering72312
Organic fine chemistryChemistry44311
Chemical engineeringChemistry44210
BiotechnologyChemistry2518
Computer technologyElectrical engineering3418
Macromolecular chemistry, polymersChemistry4206
PharmaceuticalsChemistry4116
Basic materials chemistryChemistry1416
Materials, metallurgyChemistry5005
Digital communicationElectrical engineering1315

* 2023–2024 is provisional and shown for orientation only; it is not part of the cohort comparison.

Multi-assignment: a family is counted once in every sector and field its IPC codes map to, so the sector column sums (75, 61, 27) exceed the distinct family counts (55, 46, 22). The sums are not a total and no share-of-whole chart is drawn from them.

Mechanical engineering halves between the cohorts, from 26 to 13 families, and the change sits almost entirely in one field: Machine tools, 17 families to 7. Instruments moves the other way (13 to 15), carried by Measurement, which is the largest single field in the second cohort with 12 families against 8 in the first. Chemistry and Electrical engineering are close to flat at sector level while their internal mix shifts — Biotechnology 2 to 5 and Basic materials chemistry 1 to 4, against Materials/metallurgy 5 to 0 and Pharmaceuticals 4 to 1.

IPC Subclass Profile

Technology fingerprint at IPC subclass level. IPC codes are collected from all members of each family; subclass titles come from the IPC edition named in the Methodology.

Data table: IPC subclasses with four or more families
IPC Subclass title 2013–2017 2018–2022 2023–2024* Total Change (families)
B21DWorking or processing of sheet metal or metal tubes, rods or profiles163019−13
G01NInvestigating or analysing materials by determining their chemical or physical properties49316+5
B29CShaping or joining of plastics73212−4
H04BTransmission35210+2
G01RMeasuring electric variables; measuring magnetic variables33390
C07DHeterocyclic compounds2316+1
C12NMicroorganisms or enzymes; compositions thereof1416+3
A61KPreparations for medical, dental or toiletry purposes4116−3
B01JChemical or physical processes, e.g. catalysis or colloid chemistry3126−2
B33YAdditive manufacturing22150
G06FElectric digital data processing22150
B21CManufacture of metal sheets, wire, rods, tubes, profiles3205−1
H04WWireless communication networks1214+1
B01DSeparation2114−1
H02JElectric power networks; circuit arrangements for power supply3014−3
A61PSpecific therapeutic activity of chemical compounds3014−3

* 2023–2024 is provisional and not part of the cohort comparison. No percentage changes are given: every cohort figure in this table is below 20 families and several are below 5, where a percentage would overstate a difference of one or two families.

Multi-assignment: a family carries several IPC codes and is counted once in each subclass it touches. The 2013–2017 column of this table alone sums to 59 for 55 distinct families, and the full subclass list runs further still. Subclass counts are not shares of the portfolio.

Sheet-metal working (B21D) drops from 16 families to 3 and materials investigation (G01N) rises from 4 to 9 — the two largest movements in the table and the same movement the WIPO sector view shows. G01R (electrical measurement) holds at 3 families in both cohorts and adds 3 more in 2023–2024, and C12N (microorganisms and enzymes) grows from 1 family to 4. Plastics shaping (B29C) and additive manufacturing (B33Y) remain present throughout, which places part of the manufacturing work outside the sheet-metal decline.

Inventors

Inventors named on three or more families. Spelling variants of the same name are merged by a case-, punctuation- and word-order-insensitive key, and families are then counted distinctly per inventor.

Data table: Inventors on three or more families
Inventor 2013–2017 2018–2022 2023–2024* Total
Tekkaya, A. Erman186024
Palzer, Stefan0527
Wietfeld, Christian3317
Frei, Stephan2316
Rehtanz, Christian5005
Künne, Bernd3115
Tiller, Jörg Christian4105
Gies, Soeren5005
Meya, Rickmer3104
Katzenberg, Frank3104
Böcker, Stefan0314
Brunschweiger, Andreas2103
Klika Skopic, Mateja2103
Krautwald, Tim2103
Ben Khalifa, Noomane2103
Pfost, Martin3003
Wernicke, Sebastian3003
Staupendahl, Daniel3003
Hering, Oliver0303
Thommes, Markus0213
Brümmer, Andreas0213
Ortiz Perez, Alvaro0213
Kockmann, Norbert0123
Jenau, Frank0123

* 2023–2024 is provisional and not part of the cohort comparison. Inventors named on fewer than three families are not listed. Inventor names are taken from all members of each family, so an inventor named on any member counts for that family once.

A. Erman Tekkaya is named on 18 of the 55 families in 2013–2017 and on 6 of the 46 in 2018–2022, the largest single change in the inventor record and the same direction as the decline in sheet-metal working. Of the 24 inventors with three or more families, 8 have no family in the first cohort at all (Palzer, Böcker, Hering, Thommes, Brümmer, Ortiz Perez, Kockmann, Jenau), and their subject areas — sensing, communications and process engineering — match the fields that grow in the technology profile. Eleven of the 24 are named in both cohorts, five appear only in the first; of those named in both, two hold their count across the break (Wietfeld 3 and 3, Frei 2 and 3).

Geographic Filing Strategy

Where the 123 families are present. A family counts once per patent authority, however many applications it has there, and members filed by any applicant of the family are included.

Data table: Patent authorities reached, 2013–2024
Authority 2013–2017 2018–2022 2023–2024* Total families
DE — DPMA41361996
EP — EPO3325462
WO — PCT24221157
US — USPTO810119
JP — JPO2204
CN — CNIPA1203
CA — CIPO1102
LU, FR, DK1 each001 each
IL, MX, AU, ES, KR01 each01 each
TW0011
Sum of the nine single-family authorities
(9 authority-family pairs across 7 distinct families)
3519
Distinct authorities1012516
Data table: Share of cohort families per main authority
Authority 2013–2017 (n=55) 2018–2022 (n=46) Change
DE — DPMA41 — 74.5 %36 — 78.3 %+3.8 pp
EP — EPO33 — 60.0 %25 — 54.3 %−5.7 pp
WO — PCT24 — 43.6 %22 — 47.8 %+4.2 pp
US — USPTO8 — 14.5 %10 — 21.7 %+7.2 pp
Families at DE only15 — 27.3 %17 — 37.0 %+9.7 pp

* 2023–2024 is provisional: national and regional phases of PCT applications from these filing years are largely still ahead, which is why EP and US coverage is low there. Those two years are excluded from the cohort comparison. Authority counts are non-exclusive — one family appears under every authority it reaches, so the column sums exceed the family totals.

Germany is the constant: 96 of 123 families have a DE member, and the share rises slightly between the cohorts (74.5 % to 78.3 %). Around it the international routes rearrange: the PCT share grows from 43.6 % to 47.8 % and the US share from 14.5 % to 21.7 %, while the EP share falls from 60.0 % to 54.3 %. The second cohort also reaches more distinct authorities (12 against 10), including first-time filings at KR, IL, MX, AU and ES, each with a single family. Read together with the DE-only share rising from 27.3 % to 37.0 %, the portfolio is spreading out at the top and concentrating at the bottom. All of these movements rest on differences of 2 to 7 families.

Collaboration Network

Organisations and individuals named as a second applicant on the same application as TU Dortmund.

Partners are found by joining on the appln_id of the applications TU Dortmund itself filed, not on the family: only an applicant standing on the same document counts as a partner. TU Dortmund is removed from the partner list by person_id, using the same set of person records that defines the portfolio, so no spelling variant of the university can reappear as its own partner. The list below contains no TU Dortmund record.

Data table: Co-applicants by cohort
Partner Type 2013–2017 2018–2022 2023–2024* Families
Universität Duisburg-EssenUniversity1012
Max-Planck-GesellschaftResearch0112
Karlsruher Institut für TechnologieUniversity1001
Technische Universität BerlinUniversity1001
Ernst-Moritz-Arndt-Universität GreifswaldUniversity1001
Technische Universität DresdenUniversity0101
Hochschule Hamm-LippstadtUniversity0101
Northeastern University (US)University0011
New York University in Abu Dhabi (AE)University0011
Fraunhofer-GesellschaftResearch0101
Leibniz-Institut für Analytische Wissenschaften – ISASResearch1001
ZBT GmbH – Zentrum für BrennstoffzellentechnikResearch1001
Bayer AGCompany0101
TRUMPF Werkzeugmaschinen GmbH + Co. KGCompany0101
Faurecia Emissions Control Technologies Germany GmbHCompany1001
Astrol Electronic AG (CH)Company1001
EF.Ruhr GmbHCompany1001
ISPT GmbH & Co. KGCompany1001
GDS Präzisionszerspanungs GmbHCompany0101
W2 Armaturen GmbHCompany0011
Bundesrepublik Deutschland (federal ministry)Government1001
Named individuals (4 persons on 2 families)Individual1102
Data table: Co-filing rate by cohort
Cohort Families Co-filed families Share
2013–2017551120.0 %
2018–202246715.2 %
2023–2024*22313.6 %
2013–20241232117.1 %

* 2023–2024 is provisional and not part of the cohort comparison. Two partners can sit on the same family, so the partner rows sum to more than the number of co-filed families. The Type column is assigned by mtc.berlin from the applicant name and legal form; PATSTAT carries no sector field. The four individuals are natural persons named as applicants alongside the university on two families.

Joint filing is the exception: 21 of 123 families (17.1 %) carry a second applicant, spread over 21 organisations and four individuals. Almost every partner appears exactly once. Only Universität Duisburg-Essen and the Max-Planck-Gesellschaft appear on two families each, and each of those pairs straddles two different periods rather than repeating within one. The partner mix is academic and research-heavy in both cohorts (6 of 12 partner records in 2013–2017, 4 of 8 in 2018–2022), and the only partners outside Germany are Astrol Electronic (CH) in the first cohort and, in the provisional 2023–2024 window, Northeastern University and New York University in Abu Dhabi on one shared family.

Grant Outcomes

Granted applications per office, by the cohort of the family they belong to. Application level, TU Dortmund applications only.

Grant rates for the 2018–2022 cohort are not a quality signal. Examination runs for several years after filing, so applications from 2018 onwards are still substantially pending at every office and their grant share is bound to be lower whatever the outcome eventually is. The two columns are printed side by side because the numbers exist, not because they are comparable. The same applies with even more force to the provisional 2023–2024 window, which is therefore not shown here.

Data table: Grant outcomes by office and cohort
Office 2013–2017 applications granted share 2018–2022 applications granted share*
DE — DPMA40717.5 %3612.8 %
EP — EPO372156.8 %26519.2 %
US — USPTO8675.0 %10440.0 %
Data table: Families with at least one granted member
Cohort Families With a granted member Share*
2013–2017553054.5 %
2018–2022461021.7 %
2023–202422522.7 %

* Shares for filings from 2018 onwards are depressed by examination pendency, not by outcome. WO (PCT) applications are excluded from the office table: a PCT application is not examined for grant and its granted flag in PATSTAT does not mean a patent was issued. US figures rest on 8 and 10 applications, so the percentage moves by more than 10 points with a single application. Two of the 247 applications are German translations of granted European patents (appln_kind = 'T'), i.e. validations rather than new filings at the DPMA; one falls in the 2018–2022 column and one in 2023–2024.

For the cohort that has finished examination, 21 of 37 EP applications (56.8 %) and 6 of 8 US applications were granted, against 7 of 40 at the German office (17.5 %). The gap between DE and EP is consistent with how these filings are used: 25 of the 41 families with a DE member in that cohort also have an EP or WO member, so the German application frequently serves as the priority filing for an international route rather than as the case that is pursued to grant. 15 families of the cohort never leave DE, and it is within that group that abandonment before examination would show up.

Citation Impact

Forward citations: how often later patent publications cite a publication belonging to one of the 123 families.

Data table: Eleven most cited families (two share rank 10)
# Title First filing year Citing publications
1Computer implemented method for hybrid simulation of power distribution network and associated communication network for real time applications201345
2Optical free-space transmission201433
3Method and device for the combined production of components by means of incremental sheet forming and additive methods in one clamping setup201431
4Device for three-dimensional additive printing operations, in particular for large-volume components, in particular according to the method of fused deposition molding (FDM)201427
5Method for monitoring airspace201423
6Device and method for changing filaments of different colour and/or different materials for the production of 3D printed parts and extruder with filament changing system201823
7Apparatus, used to produce hydrogen by thermal decomposition of e.g. methane, comprises pyrolysis reactor having reaction chamber present in form of capillary201321
8System and method for controlling the frequency stability and/or voltage stability of the distribution network201315
9Verfahren und Vorrichtung zur Ermittlung charakteristischer Materialkennwerte, insbesondere von Fließkurven von Leichtbauwerkstoffen, insbesondere für das Strangpressen201514
10Verfahren zur Herstellung thermoplastischer Faser-Metall-Laminat-Bauteile mittels Umformverfahren201413
10DNA-encoded chemical library, use thereof and method to synthesize the library201513
Data table: Citation coverage by cohort
Cohort Families With at least one citation Share Sum of per-family citing publications
2013–2017554174.5 %358
2018–2022462860.9 %91
2023–20242229.1 %2
2013–20241237157.7 %451

Two titles are shortened at a clause boundary for the table (ranks 7 and 10). The metric counts distinct citing patent publications per family, not citing families, and self-citations are not removed. The last column of the coverage table sums those per-family counts, so a publication citing two of these families is counted twice there; it is a sum, not a distinct total. Citation counts accumulate over time, so an older family has had more years in which to be cited; the cohort figures above are reported for completeness and cannot be read as a difference in impact between the cohorts.

The most cited family (45 citing publications, filed 2013) concerns hybrid simulation of a power distribution network together with its communication network. Four of the eleven sit in additive manufacturing and sheet forming (ranks 3, 4, 6 and 10) and three in energy and grid technology (ranks 1, 7 and 8), which are the areas the technology profile shows as the strength of the earlier cohort. The one entry from the second cohort, a filament changer for 3D printing filed in 2018, has accumulated 23 citing publications in a shorter window than any family above it.

The Current Filing Front — 2023 and 2024

Not a selection: this is the complete list of all 22 families whose earliest filing year is 2023 or 2024, in filing order. It shows what the university filed most recently, without any editorial choice.

First filing year Title
2023Cannabichromenic acid synthase variants and uses thereof
2023Compounds for use in the treatment of gastrointestinal stromal tumor (GIST)
2023Testing apparatus and method for assessing the fatigue behavior of a material on flat test pieces by means of ultrasound
2023Method for controlling a spiroergometry system
2023Shower base unit
2023Optically-based electric current metering system and use thereof
2023Sensor for determining the concentration of a gas
2023Active adaptive filter with a switch signal-based pre-control path and a switch signal delay function in the main path
2023Druckeinrichtung für den 3D-Druck
2023Vorrichtung und Verfahren zur nicht-invasiven Messung der Schmelzetemperatur in Extrusionsprozessen
2023Vorrichtung und Verfahren zur 3D-Tensor-Vervollständigung
2024Method and system for hydrogenating unsaturated hydrocarbon compounds
2024Apparatus and method for expanding gases that are under high pressure by means of an expansion machine that can be integrated into pressure control devices
2024Method for producing solids
2024Cellular user localization system: SMS side-channel timing analysis method and apparatus
2024Heizflächenelement, Heizflächeneinheit, Verfahren zum Betreiben einer Heizflächeneinheit
2024Broadband low-inductance ohmic-capacitive voltage divider for a partial discharge measurement
2024Reflector assembly for reflecting a radio signal
2024Integrated state monitoring function of a battery
2024Lichtdurchlasselement für ein Dachoberlicht
2024Greifvorrichtung zum kraftschlüssigen Greifen oder Positionieren von Bauteilen
2024Herstellung und Verwendung von kumulierten S-Yliden der Struktur R1R2S=CN2

Titles are taken from the family member with an English title where one exists, otherwise from the German original; one title is shortened at a clause boundary for the table. These 22 families are provisional data (2023–2024 filings are still being published) and are excluded from every cohort comparison in this report.

Seven of the 22 most recent filings measure or sense a physical quantity — gas concentration, melt temperature in extrusion, partial discharge, electric current, material fatigue by ultrasound, battery state and spiroergometry — which is the same area that grows in the WIPO field view (Measurement: 8 families in 2013–2017, 12 in 2018–2022, 6 in these two provisional years). Five are chemistry or biotechnology filings. The remaining ten spread across 3D printing equipment, gas expansion machinery, gripping technology, mobile-network localisation, radio reflectors, adaptive filtering, tensor completion, heating surfaces, a roof-light element and a shower base unit. Because these two years are incomplete, the list shows direction, not volume.

Methodology

Data basis, entity scope, counting rules and known limitations.

Data basis of this edition. All figures in this report were computed on 15 August 2026 against EPO PATSTAT Global, Spring 2026 (patstat-mtc.patstat_2026a) on Google BigQuery, with classification titles from IPC edition 20260101 and CPC edition 202608. Every number printed here comes from a query executed in that session; the SQL is published with the report in queries.sql.

Search strategy, entity scope and counting rules

How the entity was defined

  • The portfolio is scoped on person_name in tls206_person, never on han_name. Twelve spellings of Technische Universität Dortmund carry at least one application filed 2013–2024, from the standard forms (TECHNISCHE UNIVERSITÄT DORTMUND, Technische Universität Dortmund, TECHNISCHE UNIVERSITAET DORTMUND) through the official long form (…, Körperschaft des öffentlichen Rechts) to character damage (TECHNISCHE UNIVERSITU T DORTMUND, TECHNISCHE UNIVERSIT?T DORTMUND) and two re-orderings (UNIV DORTMUND TECH, Technical University of Dortmund). All twelve are kept. Per-spelling family counts overlap heavily, which is exactly what consolidation removes; the deduplicated portfolio is 123 families.
  • Deliberately excluded: Fachhochschule Dortmund (a separate institution; two spellings with 9 and 1 in-window families), the Leibniz-Institut für Arbeitsforschung an der TU Dortmund and the Institut für Arbeitsphysiologie (legally independent institutes, no in-window families), and the pre-2007 name Universität Dortmund (its last in-window filing year is 2011, so it contributes nothing here).
  • Office-side name renderings were searched, not assumed away. One non-Latin applicant spelling exists on the portfolio, the Chinese rendering 多特蒙德工业大学 on one family. That family is already in scope through its Latin-spelled members, so no family is reached only by a non-Latin name.

Why han_name is a cross-check and not the scope — measured on this edition

  • Contamination. Five of the twelve spellings carry han_name = TECHNISCHE UNIVERSITAT MUNCHEN. That harmonised group holds 30 distinct person_name values and 1,074 families with a 2013–2024 filing, of which 102 are TU Dortmund and 973 belong to other organisations — among them Technische Universität München itself and Technische Universität Berlin. Scoping on that group would have produced a portfolio in which 973 of 1,074 families belong to somebody else.
  • Incompleteness and fragmentation. In the opposite direction, TU Dortmund fragments across seven han_name groups in this window. The largest clean, Dortmund-specific group (TECHNISCHE UNIV DORTMUND KORPERSCHAFT DES OFFENTLICHEN RECHTS) covers 31 families; the other five clean groups hold one or two each. A scope built on any single clean group would miss the large majority of the portfolio.
  • Both directions were checked, and the person_name scope was kept.

Counting rules

  • Families. Everything is counted as DOCDB patent families (docdb_family_id), so one invention filed in several countries counts once. Pseudo-families (docdb_family_id = 0) are excluded.
  • Year of a family. A family is dated by the minimum filing year across all of its members, taken before the 2013–2024 window is applied. Taking the minimum only over members inside the window would pull families that started earlier into the first year of the window; both readings pass a sum check, so the definition has to be stated rather than tested. Consequently the annual figures add exactly to 123.
  • Two levels, kept apart. Family-level facts (first filing year, presence at an authority, any granted member) are computed over all members of the family. Application counts and grant rates by office are computed over the applications on which TU Dortmund is itself an applicant. Each table states which level it uses.
  • PCT rate is the share of families having any WO member — not the share of WO applications. Both readings were computed and are identical here (24, 22 and 11 families per cohort).
  • Co-applicants are joined on appln_id, so only an applicant on the same document counts as a partner, and TU Dortmund is removed by person_id rather than by a name pattern. A family whose members are filed separately by two organisations therefore does not count as a collaboration.
  • Rights type. The portfolio is restricted to patents of invention (ipr_type = 'PI'). This removes exactly one application from the set: a German utility model (Gebrauchsmuster) filed in 2019, whose family stays in scope through its patent applications. There are no design filings in the set. So "utility models excluded" here means one document, not a class of activity that was cut away.

Why the cohorts stop in 2022

  • The report window is 2013–2024, but the two compared cohorts are 2013–2017 and 2018–2022 — two complete five-year blocks. Filings from 2023 and 2024 are still being published into PATSTAT and their national and regional phases largely lie ahead, so a cohort ending there would measure publication lag as if it were a change in filing behaviour, in the direction that makes recent activity look weaker.
  • Equal length matters as much as completeness: nearly every table here is a count against a count, so a six-year block compared with a five-year block would show a decline created by arithmetic alone.
  • The 22 families first filed in 2023–2024 are shown in the annual trend, in a dedicated section and as a separate column in the breakdown tables, always marked provisional, and are never part of a cohort comparison. 55 + 46 + 22 = 123.

Classification scope

  • The portfolio of this report is defined by applicant name, not by a set of IPC or CPC codes. Subtree expansion, union of predecessor symbols and retention of deprecated codes therefore do not apply: no code set bounds the document set, and no patent can be lost by an incomplete one. IPC codes are used only to describe a portfolio that was already assembled.
  • Where IPC subclasses are shown, codes are collected from all members of each family, and subclass titles are joined from the classification views on symbol_patstat without any filter on is_active or on hierarchy level, so retired symbols keep their assignments and their label. IPC edition 20260101, CPC edition 202608.
  • WIPO technology fields come from tls230_appln_techn_field joined to the 35-field WIPO concordance.

What this report does not contain

  • There is no "top applicants" ranking: the applicant is the subject of the report, and every family in scope is one of its own.
  • No external benchmark is used. A comparison against a published figure for TU Dortmund could not be made because the university's transfer pages state no annual application count, so the external plausibility check was not performed rather than reported as passed.

Scope Limitations

  • Small numbers throughout. The portfolio holds 7 to 15 families per year. Most cohort differences in this report are 2 to 7 families. Percentages are printed alongside the absolute counts and never instead of them, and no percentage change is given where the base is below ten.
  • 2023 and 2024 are provisional. Applications from these filing years are still entering PATSTAT. They are shown separately and excluded from every comparison.
  • Grant rates for filings from 2018 onwards are depressed by examination pendency, not by outcome, and are not comparable with the 2013–2017 cohort. PCT applications are excluded from the grant table because they are not examined for grant.
  • Applicant-only view. Inventions where TU Dortmund researchers are named as inventors but the university is not an applicant — for instance because the rights sit with a company or with an individual — are outside this portfolio by construction.
  • Sector labels for partners are editorial. PATSTAT carries no sector field; the university, research, company, government and individual labels in the collaboration table were assigned by mtc.berlin from the applicant name and legal form.
  • Citations are not normalised. Counts are citing patent publications, self-citations included, and older families have had more time to accumulate them.
  • Filing year, not publication year. This report counts by filing year. Register searches that count by publication year will return different totals for the same institution and the same period, because a document is published about 18 months after it is filed.
Glossary — patent terms used in this report
Patent family (DOCDB)
A group of patent applications protecting the same invention in different countries. Counted once, so that international filing does not inflate the totals.
First filing year
The earliest filing year among all members of a family. Used here to place each family in exactly one year and one cohort.
DE / EP / WO / US / JP / CN / KR
Patent authority codes: DE = German Patent and Trade Mark Office, EP = European Patent Office, WO = international application under the PCT, US = United States, JP = Japan, CN = China, KR = Korea.
PCT (Patent Cooperation Treaty)
An international filing route: one application preserves the option of entering national or regional procedures in the member states for up to 30 months.
PCT rate
The share of families that have at least one WO member.
IPC
International Patent Classification, a hierarchical technology classification maintained by WIPO. A subclass is the four-character level, for example G01N.
WIPO technology fields
A concordance mapping IPC codes onto 35 technology fields in 5 sectors, used for high-level technology profiles.
han_name
PATSTAT's harmonised applicant name. It groups spellings, but it both merges different organisations and splits one organisation across several groups, so it is used in this report only as a cross-check.
Co-filing
An application carrying more than one applicant, indicating joint ownership of the invention.
Grant rate
The share of applications that were granted. Rates for recent filing years look low because examination takes several years.
Forward citation
A later patent publication citing an earlier one. Used here as a rough indicator of technical reception.

Report files

This report is published as a set of source files: meta.json with the report metadata, cover.html and kpi-strip.html, one HTML file per section under sections/, and queries.sql containing every SQL statement behind the figures, with the PATSTAT and classification editions recorded in its header. Data source: EPO PATSTAT Global, Spring 2026 (patstat-mtc.patstat_2026a) via Google BigQuery · mtc.berlin.

Like what you see?

This report was built with a fully reproducible pipeline: EPO PATSTAT Global on BigQuery, a custom MCP server, and Claude AI for analysis and visualisation. Every figure is backed by SQL that ships with the report.

Custom Patent Report Need a similar analysis for your institution, technology area or research question? We build tailored patent intelligence reports.
The Toolchain Want to run your own AI-supported patent analyses? We help you set up PATSTAT on BigQuery with MCP server integration.