As a young lad, analytical chemistry was my jam. I liked big cool machines with lasers and flames. I liked dissolving things in acid. I liked anything that gave off a swirly gas when heated.
For that reason, I have a special empathy for the plaintiffs in Harmony Biosciences, LLC v. Lupin Ltd., C.A. No. 23-1286-JLH-SRF (D. Del. Oct. 27, 2025), and their poor, overworked experts.
The patent there was for a specific crystalline form of a drug. Typically, you would figure out this form via X-ray diffraction (XRD), which involves shooting X-rays at a crystal (typically powdered) from different angles. Via science too involved to get into here, you get a series of peaks from which you can deduce how those atoms are arranged in a crystal (face-centered cubic, close-packed cubic, primitive orthorhombic, etc.).
The asserted claims specifically listed the locations of these peaks in the XRD.
It might be apparent from the above that this whole process is a real pain. It's a finicky business compared to other forms of spectroscopy, difficult to calibrate and keep steady.
So when plaintiffs received samples from all of the numerous defendants, they had a clever idea. They took the XRD of one of the samples, finding that it matched the claims.
Then, they took an NMR of that same sample. NMR operates on totally different principles, subjecting a substance to a strong magnetic field, then wiggling that field a bit, and seeing how long it takes for the various nuclei to fall back into order. This can give you broadly similar information as an XRD, but the actual output of the device is totally different. The peaks listed in the patent would not match what the NMR spits out.
Complicated as the above sounds, its actually a simpler process to do. An NMR is a giant machine (the one I used in college was about the size of a VW beetle) in a temperature controlled room that never moves. There's some initial calibration to set things up, but once you get it going you can blast through a bunch of samples.
So, plaintiffs plan was to just take an XRD of one of the samples (as it happened they use MSN's sample) to confirm it matched the claims. Then, they take NMRs of all of the samples. If the other sample's NMR's matched MSN's then one could assume that they XRD's would match as well, saving the work of doing have a dozen more XRDs.
The non-MSN defendants, however, all argued in their noninfringement reports that this doesn't work. I.e., you can't just assume that if the NMRs are the same, the XRD's would be the same as well.
So, the plaintiff went back and did an XRD on another of the samples (but not all of them)to show that they did, in fact, get the same results. Defendants moved to strike this report, arguing that this testing should have been included in an opening report, rather than a reply.
Judge Fallon agreed, striking the report and mentioning both Rule 16s good cause standard and Pennypack:
Plaintiffs first disclosed s-XRPD testing data regarding Movants' ANDA products in Parent's "reply" expert report served on October 9, 2025. (D.I. 355, Ex. A) Parent's reply report is untimely "because the expert report process as to that subject matter w[ as] never ... properly set in motion in the first place[,]" and it precludes "a direct response from any opposing expert" on his s-XRPD methodology . . . Movants aptly state that, "[i]f Plaintiffs intended on carrying their burden of proof against any Defendant via s-XRD testing, whether in addition to or instead of ssNMR testing, then they were required to include that theory and supporting evidence in their opening expert reports, not three months later."
Id. at 5 (internal citations omitted)
In conducting the Pennypack analysis, the Court noted that the burden factor favored exclusion precisely because XRD was such a pain:
Movants explain that access to specialized facilities capable of performing s-XRD testing is limited, and testing likely could not be completed in a timely fashion. Plaintiffs respond that Hikma has access to an expert who can perform s-XRPD testing, and AET and Novitium have not adequately demonstrated the need for more time to perform the testing. The court is not persuaded that Movants can perform s-XRPD testing, retain experts, and issue surreply reports in a time frame that would not disrupt the January 21, 2026 pretrial conference date, especially in the midst of expert depositions. The prejudice, cure, and disruption of trial factors therefore weigh in favor of Movants.
Id. at 10.
In closing, I am aware that no one wanted to read all this chemistry business, but it is my blog so you can't stop me (only Andrew can stop me because he's the one who understands how the website works).
If you enjoyed this post, consider subscribing to receive free e-mail updates about new posts.
