DOE is two different funders wearing one name. Most deep-tech founders have heard of ARPA-E. Far fewer have looked at the DOE Office of Science SBIR priorities, even though the Office of Science is the nation's largest supporter of basic research in the physical sciences and runs a topic-based SBIR program that fits companies ARPA-E would reject.
Here is the answer in one paragraph. The DOE Office of Science funds small businesses through topic-based SBIR/STTR solicitations tied to six research programs: Basic Energy Sciences, Advanced Scientific Computing Research, Biological and Environmental Research, Fusion Energy Sciences, High Energy Physics, and Nuclear Physics. Winning starts with matching your technology to a published research priority, not with claiming disruption.
That last sentence is the whole game. ARPA-E rewards 10x claims. The Office of Science rewards alignment. Founders who pitch one in the other's language spend 60 to 100 hours writing an application (internal Cada estimate across first-time federal applicants) that was never competitive.
This guide shows you the actual priority terms Cada extracted from Office of Science strategic documents, so you can check alignment before you spend a single hour drafting.
What is the DOE Office of Science SBIR program?
The DOE Office of Science SBIR/STTR program awards Phase I grants to small businesses whose technology advances one of its six basic-research programs. Topics are published before each solicitation and map to research program priorities. Unlike ARPA-E, which funds disruptive energy technology through its own announcements, the Office of Science funds work serving its long-horizon research mission.
A few structural facts matter here.
The Office of Science runs six research programs: Advanced Scientific Computing Research (ASCR), Basic Energy Sciences (BES), Biological and Environmental Research (BER), Fusion Energy Sciences (FES), High Energy Physics (HEP), and Nuclear Physics (NP). These six names come verbatim from DOE's own strategic plan.
The money flows through a different pipe than ARPA-E's. Office of Science financial assistance uses CFDA number 81.049; ARPA-E awards use 81.135. When we pull federal award data for clients, these are two separate assistance listings, and treating them as one funder is a category error the government's own accounting does not make.
DOE SBIR runs on a topic model. Topics are released weeks before the funding opportunity opens, typically in two cycles per year, and a letter of intent is generally required before you can submit a full application. Deadlines and amounts shift by cycle, so always verify against the current solicitation on the DOE SBIR site.
Phase I awards have recently run in the $200,000 to $250,000 range depending on topic, with Phase II awards commonly in the $1.1 million to $1.6 million range. Check the live solicitation for exact figures. The point for planning purposes: Office of Science small business funding is real money for a seed-stage deep-tech company, at roughly the same scale as an NSF SBIR.
DOE SBIR vs ARPA-E: what's the difference?
If you remember one thing: ARPA-E asks "why has nobody done this?" and the Office of Science asks "which of our priorities does this serve?" ARPA-E funds non-incremental technology where a 10 percent improvement is a rejection. Office of Science SBIR funds tools, instruments, materials, and software that advance published basic-research priorities, where alignment beats disruption.
Here is the comparison in table form.
| Dimension | Office of Science SBIR | ARPA-E |
|---|---|---|
| Core question | Does this advance a published research priority? | Is this a non-incremental leap nobody else will fund? |
| Innovation bar | Genuine advance within an established priority area | 10x metric with a credible path; incremental improvement is a rejection |
| Framing language | Core Research Areas, program priorities, user facilities | Heilmeier questions, limits of current practice, new mechanisms |
| Structure | Published topics, letter of intent, fixed cycles | Program-driven funding announcements (OPEN, focused programs) |
| Commercialization | Matters, but the research mission leads | Technology-to-market is co-equal with technical merit |
| Assistance listing | CFDA 81.049 | CFDA 81.135 |
The ARPA-E column is not speculation. Cada maintains a full ARPA-E concept-paper playbook, and that playbook's review rubric caps "engineering optimization of a known approach" at a rejection-level score no matter how well written the pitch is.
ARPA-E reviewers weight disruptive innovation at 25 percent of the total evaluation. There is no equivalent disqualifier in Office of Science SBIR review, where fit to the topic is the organizing test.
A fictional example makes the difference concrete. Imagine a startup with a new electrolyte additive that extends lithium-ion cycle life by 15 percent.
Pitched to ARPA-E: dead on arrival. A 15 percent improvement through formulation chemistry is engineering optimization, and ARPA-E's own evaluation lens classifies that as a rejection regardless of market size.
Pitched to Office of Science BES: viable, if reframed. BES lists "synthesis and processing science" and "condensed phase and interfacial molecular science" among its Core Research Areas. An SBIR topic descended from those priorities wants better interfacial chemistry understanding and the instruments and materials that enable it. The same technology, framed as advancing interfacial science with a commercial product as the vehicle, is now speaking the reviewer's language.
Same company, same molecules, opposite outcomes. That is what "two funders wearing one name" means in practice.
How does the DOE Office of Science pick SBIR topics?
Topics are not invented fresh each cycle. They descend from the research programs' standing priorities, which are published in strategic documents, Core Research Area lists, and advisory committee reports.
This gives you a reading order that most applicants skip:
- Read the program's priority documents first (strategic plans, research pages, advisory reports).
- Map your technology to the specific priority terms those documents use.
- Only then read the current topic list, looking for topics that descend from the priorities you matched.
Founders usually do this backwards. If you are trying to figure out how to choose a DOE SBIR topic, the honest answer is that you choose a priority first and let the topic follow. Finding a vaguely relevant topic and stretching your technology to fit it produces an application that reads as stretched because it is.
Reading priorities first has a second benefit: topics rotate, priorities persist. If you match a durable priority and miss this cycle's topic, you are positioned for the next cycle instead of starting over.
The average founder we work with spends 60 to 100 hours on a first federal application. Spending the first 2 of those hours on priority documents is the cheapest risk reduction available.
DOE Office of Science SBIR priorities: the Basic Energy Sciences list
BES is the natural home for materials, chemistry, and instrumentation companies. It organizes its research portfolio into named Core Research Areas (CRAs), and those exact names are the highest-value vocabulary in a BES-aligned application.
This is also why searching for DOE Basic Energy Sciences SBIR topics without this vocabulary underperforms. The topics are written in CRA language, so knowing the CRA names is what makes the topic list legible.
Cada extracted the following priority terms from BES research pages. They are drawn directly from DOE's own published materials, and a human reviewer verified the extraction against DOE's source pages in April 2026.
BES Core Research Areas and priority terms:
| Priority term | Category |
|---|---|
| Materials chemistry | Core Research Area |
| Biomolecular materials | Core Research Area |
| Synthesis and processing science | Core Research Area |
| Experimental condensed matter physics | Core Research Area |
| Theoretical condensed matter physics | Core Research Area |
| Solar photochemistry | Core Research Area |
| Photosynthetic systems | Core Research Area |
| Physical biosciences | Core Research Area |
| Catalysis science | Core Research Area |
| Heavy element chemistry | Core Research Area |
| Separation science | Core Research Area |
| Geosciences research | Core Research Area |
| Gas phase chemical physics | Core Research Area |
| Condensed phase and interfacial molecular science | Core Research Area |
| Atomic, molecular, and optical sciences | Core Research Area |
| Mechanical behavior and radiation effects | Core Research Area |
| Physical behavior of materials | Core Research Area |
| Computational and theoretical chemistry | Core Research Area |
| X-ray scattering | Technology domain |
| Neutron scattering | Technology domain |
| Electron and scanning probe microscopies | Technology domain |
| Accelerator and detector research | Technology domain |
Two named programs are worth knowing even as a small business. Energy Frontier Research Centers drew a single funding announcement of $352 million, per DOE's published figure, which tells you where BES concentrates long-horizon bets. Nanoscale Science Research Centers are user facilities, and SBIR topics regularly ask for instruments and capabilities that serve facility users.
If you build scientific instruments, this table should change how you see DOE. X-ray scattering, neutron scattering, and electron microscopy are not niche keywords. They are standing technology domains with facility ecosystems behind them and recurring small-business needs.
ASCR priorities: computing, AI, and quantum
ASCR is the research program most founders do not know exists. If you build computing hardware, scientific software, or AI tools, DOE may be a fit even though nothing you make plugs into a wall.
From Cada's extraction of ASCR research priorities:
- Scalable algorithms and mathematics libraries
- Multiscale and multi-physics modeling
- Scientific AI (the term appears repeatedly in ASCR's own documents)
- Efficient analysis of massive data sets
- High-performance computing and scientific-data infrastructure
- Quantum computing and networking, and quantum information systems
- Neuromorphic computing
- Robotics and microelectronics
- Cybersecurity investments specific to scientific computing
Named programs matter here too. SciDAC (Scientific Discovery through Advanced Computing) signals that DOE pays for software that makes science faster. The Lab Embedded Entrepreneurship Program (LEEP) embeds founders inside national labs, which is a signal about how DOE thinks small companies and labs should connect.
A fictional example: a three-person startup building GPU-accelerated sparse linear algebra libraries has no obvious energy story. It maps directly to "scalable algorithms and mathematics libraries." That company should be reading ASCR-descended SBIR topics before it reads anything at ARPA-E.
BER priorities: read the right document
Honesty requires a caution here. If you have searched for DOE BER SBIR funding priorities and come away confused, there is a structural reason.
When we ran our extraction on BER's public community-resources page, the highest-ranked terms included wildfire plume modeling and a repurposed hepatitis C drug for COVID treatment. Those are research news highlights, not durable program priorities.
That is a data-quality lesson we would rather teach than hide: not every agency page is equally authoritative. BES publishes a structured Core Research Area list; BER's public-facing page leads with news. Frankly, our structured BER priority data is thinner than our BES and ASCR data because of exactly this.
What the BER extraction does establish: earth system modeling is a live technology domain, and BERAC (BER's advisory committee), BRaVE (Biopreparedness Research Virtual Environment), and NVBL (National Virtual Biotechnology Laboratory) are named structures worth knowing.
Practical guidance: for BER, weight advisory committee reports (BERAC) and the current SBIR topic descriptions more heavily than the program's news page. If your technology sits in environmental genomics, earth systems, or bio-instrumentation, the topics themselves are the more reliable priority signal for this program.
How to use DOE Office of Science SBIR priorities in your application
This is the part of the methodology we normally apply inside client engagements. Here is the full version.
Step 1: Run the alignment test. Put the priority-term table for your target program next to a one-paragraph description of your technology. If no priority term describes what you do, stop. You are at the wrong research program or the wrong agency entirely. No amount of writing quality fixes a missing priority match, and this test takes 30 minutes instead of 80 hours.
Step 2: Mirror the program's exact vocabulary in your significance section. If BES says "condensed phase and interfacial molecular science" and you write "we study wet surfaces," the reviewer does the translation work, and reviewers do not reliably do translation work. Use the agency's own terms to describe where your work sits, then use your own terms to describe what is new about it.
Step 3: Name the priority your technical approach advances. Somewhere in your approach section, one sentence should say, in effect: this work advances [Core Research Area] by [specific contribution]. This is the sentence a program manager can quote when defending your award internally.
Three steps, and none of them are writing tricks. They are fit checks. The strongest DOE applications we have seen were competitive at step 1, before a single section was drafted.
One more honest note: mirroring vocabulary does not mean keyword stuffing. Reviewers are scientists in the field. One precise use of "separation science" placed where it genuinely applies beats five decorative uses.
Which DOE door should you knock on?
Use your technology's posture, not its sector, to route:
| Your situation | Best DOE door |
|---|---|
| Genuine advance inside an established research priority (better instrument, better material, better algorithm) | Office of Science SBIR, topic-matched |
| New mechanism with a credible 10x claim against a named limit of current technology | ARPA-E |
| Near-commercial technology needing deployment scale, not research | DOE applied offices (efficiency, grid, manufacturing programs) and their SBIR topics |
| Computing or scientific software with no energy story | Office of Science, ASCR-descended topics |
The routing mistake we see most is the middle case in reverse: a company with a solid 15 to 30 percent improvement pitching ARPA-E because ARPA-E is the DOE brand they know. ARPA-E's own evaluation framework classifies that as engineering optimization and rejects it. The same company often has a legitimate Office of Science or applied-office path it never examined.
Not sure which DOE program fits?
You now have the actual priority vocabulary for DOE Office of Science SBIR priorities across BES and ASCR, the caution flag on BER, and the routing logic that separates Office of Science from ARPA-E. The 30-minute version of this work is the alignment test in step 1 above, and you can run it yourself with the tables on this page.
If you want a second opinion, Cada runs a free 15-minute DOE program routing call: you describe the technology, we tell you straight whether it fits Office of Science, ARPA-E, or an applied office, and which priority language matches. No pitch, no obligation, nothing to pay. Cada has written 100+ proposals across 30+ agencies, and the fastest way we save founders money is telling them which applications not to write.
Sources
- DOE Office of Science SBIR/STTR program -- topic-based solicitation structure, letter-of-intent requirement, and current topic lists
- DOE Office of Science -- the six research programs and the basic-research mission statement
- Basic Energy Sciences -- Core Research Area names in the BES priority table, Energy Frontier Research Centers, and Nanoscale Science Research Centers
- Advanced Scientific Computing Research -- ASCR research priorities, SciDAC, and LEEP
- Biological and Environmental Research -- BER program structure, BERAC, BRaVE, and NVBL
- ARPA-E -- program announcements and the evaluation approach referenced in the comparison table
- SAM.gov Assistance Listings -- CFDA 81.049 (Office of Science) and 81.135 (ARPA-E)
- Priority-term extractions, the ARPA-E review-rubric characterization, and the 60-to-100-hour first-application estimate are Cada's own data, drawn from our agency intelligence database and client work (100+ proposals across 30+ agencies)
Award amounts, cycle timing, and letter-of-intent requirements change by solicitation. Verify everything against the live solicitation on the DOE SBIR site before you plan around it. All company examples are fictional and used for illustration only.